diff --git a/.gitattributes b/.gitattributes index 660eddc40fdd01bbb6cae296dc0369ae49e9ef5d..e47bcd11033628313d948ac5bb46f323afa734b4 100644 --- a/.gitattributes +++ b/.gitattributes @@ -246,3 +246,9 @@ logs/13849784-bench_1.34G_dp2_tp16_pp16_acc32_mbs4_seq4096_zero1_tpmodeRED_vocab logs/13850047-bench_1.34G_dp32_tp8_pp2_acc8_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out filter=lfs diff=lfs merge=lfs -text logs/13849388-bench_1.34G_dp2_tp8_pp16_acc64_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out filter=lfs diff=lfs merge=lfs -text logs/13850064-bench_469G_dp32_tp8_pp2_acc4_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out filter=lfs diff=lfs merge=lfs -text +logs/14099259-bench_3.57G_dp8_tp4_pp8_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out filter=lfs diff=lfs merge=lfs -text +logs/13849751-bench_80G_dp64_tp2_pp2_acc4_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out filter=lfs diff=lfs merge=lfs -text +logs/14099174-bench_1.34G_dp16_tp8_pp2_acc1_mbs16_seq4096_zero1_tpmodeRED_vocab131k.out filter=lfs diff=lfs merge=lfs -text +logs/13849583-bench_1.34G_dp8_tp4_pp8_acc16_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out filter=lfs diff=lfs merge=lfs -text +logs/13849976-bench_469G_dp8_tp32_pp2_acc1_mbs32_seq4096_zero1_tpmodeRED_vocab131k.out filter=lfs diff=lfs merge=lfs -text +logs/13849695-bench_80G_dp16_tp8_pp2_acc2_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out filter=lfs diff=lfs merge=lfs -text diff --git a/logs/12346536-bench_3.56G_dp16_tp4_pp1_acc2_mbs16_seq4096_zero1_l28_h3072_heads24.out b/logs/12346536-bench_3.56G_dp16_tp4_pp1_acc2_mbs16_seq4096_zero1_l28_h3072_heads24.out new file mode 100644 index 0000000000000000000000000000000000000000..940d784fbaf2cc165c90f9b3fe8c3ad8baf0dc9e --- /dev/null +++ b/logs/12346536-bench_3.56G_dp16_tp4_pp1_acc2_mbs16_seq4096_zero1_l28_h3072_heads24.out @@ -0,0 +1,1984 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-165-131,ip-26-0-166-214,ip-26-0-168-52,ip-26-0-170-[143,160],ip-26-0-171-168,ip-26-0-172-[73,116]' ++ export 'NODELIST=ip-26-0-165-131 +ip-26-0-166-214 +ip-26-0-168-52 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-168 +ip-26-0-172-73 +ip-26-0-172-116' ++ NODELIST='ip-26-0-165-131 +ip-26-0-166-214 +ip-26-0-168-52 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-168 +ip-26-0-172-73 +ip-26-0-172-116' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-165-131,ip-26-0-166-214,ip-26-0-168-52,ip-26-0-170-[143,160],ip-26-0-171-168,ip-26-0-172-[73,116]' ++ export MASTER_NODE=ip-26-0-165-131 ++ MASTER_NODE=ip-26-0-165-131 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=8 ++ NNODES=8 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=64 ++ WORLD_SIZE=64 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-165-131' +Master node: ip-26-0-165-131 ++ echo 'All nodes: ip-26-0-165-131 +ip-26-0-166-214 +ip-26-0-168-52 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-168 +ip-26-0-172-73 +ip-26-0-172-116' +All nodes: ip-26-0-165-131 +ip-26-0-166-214 +ip-26-0-168-52 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-168 +ip-26-0-172-73 +ip-26-0-172-116 ++ echo 'World size: 64' +World size: 64 ++ srun torchrun --nnodes=8 --nproc_per_node=8 --rdzv_id=12346536 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-165-131:12356 run_train.py --config-file benchmark/configs/config_3.56G_dp16_tp4_pp1_acc2_mbs16_seq4096_zero1_l28_h3072_heads24.yaml +[2024-12-03 10:30:09,423] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-03 10:30:09,423] torch.distributed.run: [WARNING] +[2024-12-03 10:30:09,423] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,423] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-03 10:30:09,423] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,437] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-03 10:30:09,438] torch.distributed.run: [WARNING] +[2024-12-03 10:30:09,438] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,438] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-03 10:30:09,438] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,438] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-03 10:30:09,438] torch.distributed.run: [WARNING] +[2024-12-03 10:30:09,438] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,438] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-03 10:30:09,438] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,450] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-03 10:30:09,450] torch.distributed.run: [WARNING] +[2024-12-03 10:30:09,450] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,450] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-03 10:30:09,450] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,455] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-03 10:30:09,455] torch.distributed.run: [WARNING] +[2024-12-03 10:30:09,455] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,455] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-03 10:30:09,455] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,494] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-03 10:30:09,501] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-03 10:30:09,519] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-03 10:30:09,494] torch.distributed.run: [WARNING] +[2024-12-03 10:30:09,494] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,494] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-03 10:30:09,494] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,501] torch.distributed.run: [WARNING] +[2024-12-03 10:30:09,501] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,501] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-03 10:30:09,501] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,519] torch.distributed.run: [WARNING] +[2024-12-03 10:30:09,519] torch.distributed.run: [WARNING] ***************************************** +[2024-12-03 10:30:09,519] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-03 10:30:09,519] torch.distributed.run: [WARNING] ***************************************** +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Config: +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Config(general=GeneralArgs(project='debug', +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: run='3.56G_dp16_tp4_pp1_acc2_mbs16_seq4096_zero1_l28_h3072_heads24', +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: seed=42, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: step=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: consumed_train_samples=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: benchmark_csv_path=PosixPath('bench_elie_4k.csv'), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: ignore_sanity_checks=True), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: parallelism=ParallelismArgs(dp=16, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: pp=1, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: tp=4, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: pp_engine=, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: tp_mode=, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: tp_linear_async_communication=True, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: recompute_layer=False, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: tp_recompute_allgather=True, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: expert_parallel_size=1), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: eos_token_id=0, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: hidden_act='silu', +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: hidden_size=3072, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: initializer_range=0.02, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: intermediate_size=8192, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: is_llama_config=True, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: max_position_embeddings=4096, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: num_attention_heads=24, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: num_hidden_layers=28, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: num_key_value_heads=8, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: pad_token_id=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: pretraining_tp=1, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: rms_norm_eps=1e-05, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: rope_scaling=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: rope_theta=10000.0, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: rope_interleaved=False, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: tie_word_embeddings=True, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: use_cache=True, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: vocab_size=128256), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: init_method=RandomInit(std=0.02), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: dtype=torch.bfloat16, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: make_vocab_size_divisible_by=1, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: ddp_bucket_cap_mb=25), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: tokenizer_revision=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: tokenizer_max_length=None), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: checkpoint_interval=10000, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: save_initial_state=False, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: save_final_state=False, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: resume_checkpoint_path=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: checkpoints_path_is_shared_file_system=False), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: logging=LoggingArgs(log_level='info', +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: log_level_replica='info', +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: iteration_step_info_interval=1), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: tokens=TokensArgs(sequence_length=4096, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: train_steps=100, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: micro_batch_size=16, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: batch_accumulation_per_replica=2, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: val_check_interval=100, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: limit_val_batches=0, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: limit_test_batches=0), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: adam_beta1=0.9, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: adam_beta2=0.95, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: torch_adam_is_fused=True, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: name='adamW'), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: zero_stage=1, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: weight_decay=0.01, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: clip_grad=1.0, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: accumulate_grad_in_fp32=True, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: lr_warmup_steps=2, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: lr_warmup_style='linear', +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: lr_decay_style='cosine', +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: lr_decay_steps=13, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: lr_decay_starting_step=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: min_decay_lr=1e-05)), +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: start_training_step=1, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: data=DataArgs(dataset=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: seed=42, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: num_loading_workers=1))], +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: profiler=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: lighteval=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: s3_upload=None) +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Model Config: +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: LlamaConfig(bos_token_id=0, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: eos_token_id=0, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: hidden_act='silu', +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: hidden_size=3072, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: initializer_range=0.02, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: intermediate_size=8192, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: is_llama_config=True, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: max_position_embeddings=4096, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: num_attention_heads=24, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: num_hidden_layers=28, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: num_key_value_heads=8, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: pad_token_id=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: pretraining_tp=1, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: rms_norm_eps=1e-05, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: rope_scaling=None, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: rope_theta=10000.0, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: rope_interleaved=False, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: tie_word_embeddings=True, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: use_cache=True, +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: vocab_size=128256) +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Building model.. +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Initialize RoPE Theta = 10000.0 +12/03/2024 10:30:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Setting PP block ranks... +12/03/2024 10:30:54 [INFO|DP=0|PP=0|TP=2|ip-26-0-165-131]: Local number of parameters: 803M (1532.21MiB) +12/03/2024 10:30:54 [INFO|DP=0|PP=0|TP=3|ip-26-0-165-131]: Local number of parameters: 803M (1532.21MiB) +12/03/2024 10:30:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Total number of parameters: 3.21G (6128.84MiB) +12/03/2024 10:30:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Local number of parameters: 803M (1532.21MiB) +12/03/2024 10:30:54 [INFO|DP=0|PP=0|TP=1|ip-26-0-165-131]: Local number of parameters: 803M (1532.21MiB) +12/03/2024 10:30:54 [INFO|DP=0|PP=0|TP=3|ip-26-0-165-131]: [After model building] Memory usage: 1533.37MiB. Peak allocated: 5408.00MiB Peak reserved: 35106.00MiB +12/03/2024 10:30:54 [INFO|DP=0|PP=0|TP=2|ip-26-0-165-131]: [After model building] Memory usage: 1533.37MiB. Peak allocated: 5408.00MiB Peak reserved: 35106.00MiB +12/03/2024 10:30:54 [INFO|DP=0|PP=0|TP=1|ip-26-0-165-131]: [After model building] Memory usage: 1533.37MiB. Peak allocated: 5408.00MiB Peak reserved: 35106.00MiB +12/03/2024 10:30:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [After model building] Memory usage: 1533.37MiB. Peak allocated: 5408.00MiB Peak reserved: 35106.00MiB +12/03/2024 10:30:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: No checkpoint path provided. +12/03/2024 10:30:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Parametrizing model parameters using StandardParametrizator +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [Optimizer Building] Using LearningRateForSP as learning rate +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] Size of optimizer params per rank: +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 0 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 1 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 2 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 3 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 4 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 5 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 6 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 7 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 8 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 9 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 10 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 11 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 12 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 13 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 14 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [ZeRO sharding] DP Rank 15 has 50.2M out of 803M (6.25%) params' optimizer states +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Using dummy data generator +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [Training Plan] There are 1 training stages +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [Stage Stable Training Stage] start from step 1 +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: +12/03/2024 10:30:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: [Start training] datetime: 2024-12-03 10:30:57.701277 | mbs: 16 | grad_accum: 2 | global_batch_size: 512 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1017, in forward + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 7.83 GiB. GPU 4 has a total capacty of 79.33 GiB of which 7.71 GiB is free. Including non-PyTorch memory, this process has 71.61 GiB memory in use. Of the allocated memory 58.68 GiB is allocated by PyTorch, and 1.81 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1017, in forward + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 7.83 GiB. GPU 0 has a total capacty of 79.33 GiB of which 7.77 GiB is free. Including non-PyTorch memory, this process has 71.55 GiB memory in use. Of the allocated memory 58.68 GiB is allocated by PyTorch, and 1.68 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1017, in forward + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 7.83 GiB. GPU 0 has a total capacty of 79.33 GiB of which 7.82 GiB is free. Including non-PyTorch memory, this process has 71.50 GiB memory in use. Of the allocated memory 58.68 GiB is allocated by PyTorch, and 1.74 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + result = loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + result = loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + grad_tensor = grad_output.matmul(weight) +RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` + grad_tensor = grad_output.matmul(weight) +RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + result = loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return user_fn(self, *args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 821, in backward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + dx, dw, db, dresidual_in, dx1, dw1, db1 = _layer_norm_bwd( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 643, in _layer_norm_bwd + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + _layer_norm_bwd_kernel[grid]( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 100, in run + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + result = loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + timings = {config: self._bench(*args, config=config, **kwargs) + torch.autograd.backward( File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 100, in + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + timings = {config: self._bench(*args, config=config, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 83, in _bench + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return do_bench(kernel_call, warmup=self.warmup, rep=self.rep, quantiles=(0.5, 0.2, 0.8)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/testing.py", line 104, in do_bench + return user_fn(self, *args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 821, in backward + dx, dw, db, dresidual_in, dx1, dw1, db1 = _layer_norm_bwd( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 643, in _layer_norm_bwd + _layer_norm_bwd_kernel[grid]( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 100, in run + fn() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 81, in kernel_call + timings = {config: self._bench(*args, config=config, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 100, in + self.fn.run(*args, num_warps=config.num_warps, num_stages=config.num_stages, **current) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 232, in run + timings = {config: self._bench(*args, config=config, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 83, in _bench + return do_bench(kernel_call, warmup=self.warmup, rep=self.rep, quantiles=(0.5, 0.2, 0.8)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/testing.py", line 104, in do_bench + return self.fn.run(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 232, in run + fn() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 81, in kernel_call + return self.fn.run(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 232, in run + self.fn.run(*args, num_warps=config.num_warps, num_stages=config.num_stages, **current) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 232, in run + return self.fn.run(*args, **kwargs) + [Previous line repeated 2 more times] + File "", line 65, in _layer_norm_bwd_kernel + return self.fn.run(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 232, in run + return self.fn.run(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/triton/runtime/autotuner.py", line 232, in run + return self.fn.run(*args, **kwargs) + [Previous line repeated 2 more times] + File "", line 65, in _layer_norm_bwd_kernel +RuntimeError: Triton Error [CUDA]: out of memory +RuntimeError: Triton Error [CUDA]: out of memory +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + result = loss.backward()result = loss.backward() + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + result = loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( +torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) +return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + grad_tensor = grad_output.matmul(weight)grad_tensor = grad_output.matmul(weight) + +RuntimeErrorRuntimeError: : CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)`CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` + + grad_tensor = grad_output.matmul(weight) +RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train +trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + self.backward(context=context, state=state, grad_accumulator=grad_accumulator)self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + result = loss.backward() + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward +grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + result = loss.backward()result = loss.backward() + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + result = loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + torch.autograd.backward(torch.autograd.backward( + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + grad_tensor = grad_output.matmul(weight) +RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward passVariable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args)return user_fn(self, *args) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + grad_tensor = grad_output.matmul(weight)grad_tensor = grad_output.matmul(weight) + +RuntimeErrorRuntimeError: : CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)`CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` + + grad_tensor = grad_output.matmul(weight) +RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + result = loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + result = loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + grad_tensor = grad_output.matmul(weight) +RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + grad_tensor = grad_output.matmul(weight) +RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + result = loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + result = loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + grad_tensor = grad_output.matmul(weight) +RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` + grad_tensor = grad_output.matmul(weight) +RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 295, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 86, in backward + grad_accumulator.backward(sum(activations)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 205, in backward + result = loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 296, in backward + grad_tensor = grad_output.matmul(weight) +RuntimeError: CUDA error: CUBLAS_STATUS_ALLOC_FAILED when calling `cublasCreate(handle)` +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train +trainer.train(dataloader) +trainer.train(dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model)output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch)output = model(**micro_batch) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0]return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0]return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"]fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + output = self.pp_block(**new_kwargs) +output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), + module_builder=lambda: lambda x: x.float(), +torch.cuda .module_builder=lambda: lambda x: x.float(),OutOfMemoryError: +CUDA out of memory. Tried to allocate 7.83 GiB. GPU 7 has a total capacty of 79.33 GiB of which 7.49 GiB is free. Including non-PyTorch memory, this process has 71.83 GiB memory in use. Of the allocated memory 55.17 GiB is allocated by PyTorch, and 6.21 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFtorch.cuda +.OutOfMemoryErrortorch.cuda: .CUDA out of memory. Tried to allocate 7.83 GiB. GPU 5 has a total capacty of 79.33 GiB of which 7.21 GiB is free. Including non-PyTorch memory, this process has 72.10 GiB memory in use. Of the allocated memory 55.17 GiB is allocated by PyTorch, and 6.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFOutOfMemoryError +: CUDA out of memory. Tried to allocate 7.83 GiB. GPU 6 has a total capacty of 79.33 GiB of which 7.03 GiB is free. Including non-PyTorch memory, this process has 72.29 GiB memory in use. Of the allocated memory 55.17 GiB is allocated by PyTorch, and 6.24 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 7.83 GiB. GPU 4 has a total capacty of 79.33 GiB of which 7.34 GiB is free. Including non-PyTorch memory, this process has 71.98 GiB memory in use. Of the allocated memory 55.17 GiB is allocated by PyTorch, and 5.96 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader)trainer.train(dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model)output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch)output = model(**micro_batch) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0]return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + output = self.pp_block(**new_kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), +module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: torch.cudaCUDA out of memory. Tried to allocate 7.83 GiB. GPU 5 has a total capacty of 79.33 GiB of which 7.00 GiB is free. Including non-PyTorch memory, this process has 72.32 GiB memory in use. Of the allocated memory 55.17 GiB is allocated by PyTorch, and 6.28 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +.OutOfMemoryError: CUDA out of memory. Tried to allocate 7.83 GiB. GPU 7 has a total capacty of 79.33 GiB of which 7.18 GiB is free. Including non-PyTorch memory, this process has 72.14 GiB memory in use. Of the allocated memory 55.17 GiB is allocated by PyTorch, and 6.53 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 7.83 GiB. GPU 4 has a total capacty of 79.33 GiB of which 6.75 GiB is free. Including non-PyTorch memory, this process has 72.57 GiB memory in use. Of the allocated memory 55.17 GiB is allocated by PyTorch, and 6.56 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 7.83 GiB. GPU 3 has a total capacty of 79.33 GiB of which 6.89 GiB is free. Including non-PyTorch memory, this process has 72.42 GiB memory in use. Of the allocated memory 55.17 GiB is allocated by PyTorch, and 6.49 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 237, in + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 490, in training_step + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 7.83 GiB. GPU 0 has a total capacty of 79.33 GiB of which 7.57 GiB is free. Including non-PyTorch memory, this process has 71.74 GiB memory in use. Of the allocated memory 55.17 GiB is allocated by PyTorch, and 5.78 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 7.83 GiB. GPU 1 has a total capacty of 79.33 GiB of which 6.56 GiB is free. Including non-PyTorch memory, this process has 72.76 GiB memory in use. Of the allocated memory 55.17 GiB is allocated by PyTorch, and 6.71 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 7.83 GiB. GPU 2 has a total capacty of 79.33 GiB of which 6.65 GiB is free. Including non-PyTorch memory, this process has 72.67 GiB memory in use. Of the allocated memory 55.17 GiB is allocated by PyTorch, and 6.62 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/03/2024 10:31:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/03/2024 10:31:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-131]: Memory usage: 4789.31MiB. Peak allocated 5408.00MiB. Peak reserved: 38172.00MiB +[2024-12-03 10:31:05,792] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 489171 closing signal SIGTERM +[2024-12-03 10:31:05,792] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 489174 closing signal SIGTERM +[2024-12-03 10:31:05,792] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 489176 closing signal SIGTERM +[2024-12-03 10:31:05,792] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 489177 closing signal SIGTERM +[2024-12-03 10:31:05,792] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 489178 closing signal SIGTERM +[2024-12-03 10:31:05,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247784 closing signal SIGTERM +[2024-12-03 10:31:05,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247785 closing signal SIGTERM +[2024-12-03 10:31:05,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247786 closing signal SIGTERM +[2024-12-03 10:31:05,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247788 closing signal SIGTERM +[2024-12-03 10:31:05,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247790 closing signal SIGTERM +[2024-12-03 10:31:05,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 844871 closing signal SIGTERM +[2024-12-03 10:31:05,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 844872 closing signal SIGTERM +[2024-12-03 10:31:05,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 844873 closing signal SIGTERM +[2024-12-03 10:31:05,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 844875 closing signal SIGTERM +[2024-12-03 10:31:05,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 844876 closing signal SIGTERM +[2024-12-03 10:31:05,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 844877 closing signal SIGTERM +[2024-12-03 10:31:05,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3198065 closing signal SIGTERM +[2024-12-03 10:31:05,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3198066 closing signal SIGTERM +[2024-12-03 10:31:05,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3198068 closing signal SIGTERM +[2024-12-03 10:31:05,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3198072 closing signal SIGTERM +[2024-12-03 10:31:05,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 119127 closing signal SIGTERM +[2024-12-03 10:31:05,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 119129 closing signal SIGTERM +[2024-12-03 10:31:05,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 119131 closing signal SIGTERM +[2024-12-03 10:31:05,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 119133 closing signal SIGTERM +[2024-12-03 10:31:05,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2825897 closing signal SIGTERM +[2024-12-03 10:31:05,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2825898 closing signal SIGTERM +[2024-12-03 10:31:05,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2825899 closing signal SIGTERM +[2024-12-03 10:31:05,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2825900 closing signal SIGTERM +[2024-12-03 10:31:05,805] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 88350 closing signal SIGTERM +[2024-12-03 10:31:05,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 88351 closing signal SIGTERM +[2024-12-03 10:31:05,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 88353 closing signal SIGTERM +[2024-12-03 10:31:05,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 88356 closing signal SIGTERM +[2024-12-03 10:31:05,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 88357 closing signal SIGTERM +[2024-12-03 10:31:09,536] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 4 (pid: 2825901) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-03_10:31:05 + host : ip-26-0-165-131.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 2825902) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-03_10:31:05 + host : ip-26-0-165-131.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 2825903) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-03_10:31:05 + host : ip-26-0-165-131.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 2825904) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-03_10:31:05 + host : ip-26-0-165-131.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 2825901) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-03 10:31:10,548] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-73.ec2.internal_388537_0' has failed to send a keep-alive heartbeat to the rendezvous '12346536' due to an error of type RendezvousConnectionError. +[2024-12-03 10:31:10,584] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-168.ec2.internal_88344_0' has failed to send a keep-alive heartbeat to the rendezvous '12346536' due to an error of type RendezvousConnectionError. +[2024-12-03 10:31:10,658] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-170-143.ec2.internal_247776_0' has failed to send a keep-alive heartbeat to the rendezvous '12346536' due to an error of type RendezvousConnectionError. +[2024-12-03 10:31:10,697] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-116.ec2.internal_844863_0' has failed to send a keep-alive heartbeat to the rendezvous '12346536' due to an error of type RendezvousConnectionError. +[2024-12-03 10:31:10,753] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-214.ec2.internal_119120_0' has failed to send a keep-alive heartbeat to the rendezvous '12346536' due to an error of type RendezvousConnectionError. +[2024-12-03 10:31:10,767] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-170-160.ec2.internal_489165_0' has failed to send a keep-alive heartbeat to the rendezvous '12346536' due to an error of type RendezvousConnectionError. +[2024-12-03 10:31:10,775] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-52.ec2.internal_3198059_0' has failed to send a keep-alive heartbeat to the rendezvous '12346536' due to an error of type RendezvousConnectionError. +[2024-12-03 10:31:10,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 388550 closing signal SIGTERM +[2024-12-03 10:31:10,825] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 119126) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-03 10:31:10,839] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 2 (pid: 3198067) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-03 10:31:10,869] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-214.ec2.internal_119120_0' has failed to shutdown the rendezvous '12346536' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-03_10:31:05 + host : ip-26-0-166-214.ec2.internal + rank : 10 (local_rank: 2) + exitcode : 1 (pid: 119128) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-03_10:31:05 + host : ip-26-0-166-214.ec2.internal + rank : 12 (local_rank: 4) + exitcode : 1 (pid: 119130) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-03_10:31:05 + host : ip-26-0-166-214.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 119132) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-03_10:31:05 + host : ip-26-0-166-214.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 119126) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-03 10:31:10,881] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-52.ec2.internal_3198059_0' has failed to shutdown the rendezvous '12346536' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-03_10:31:05 + host : ip-26-0-168-52.ec2.internal + rank : 20 (local_rank: 4) + exitcode : 1 (pid: 3198069) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-03_10:31:05 + host : ip-26-0-168-52.ec2.internal + rank : 21 (local_rank: 5) + exitcode : 1 (pid: 3198070) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-03_10:31:05 + host : ip-26-0-168-52.ec2.internal + rank : 22 (local_rank: 6) + exitcode : 1 (pid: 3198071) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-03_10:31:05 + host : ip-26-0-168-52.ec2.internal + rank : 18 (local_rank: 2) + exitcode : 1 (pid: 3198067) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-165-131: task 0: Exited with exit code 1 +srun: error: ip-26-0-168-52: task 2: Exited with exit code 1 +srun: error: ip-26-0-166-214: task 1: Exited with exit code 1 +[2024-12-03 10:31:11,938] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 247783) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-03 10:31:11,978] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-170-143.ec2.internal_247776_0' has failed to shutdown the rendezvous '12346536' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-03_10:31:05 + host : ip-26-0-170-143.ec2.internal + rank : 28 (local_rank: 4) + exitcode : 1 (pid: 247787) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-03_10:31:05 + host : ip-26-0-170-143.ec2.internal + rank : 30 (local_rank: 6) + exitcode : 1 (pid: 247789) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-03_10:31:05 + host : ip-26-0-170-143.ec2.internal + rank : 24 (local_rank: 0) + exitcode : 1 (pid: 247783) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-03 10:31:12,094] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 2 (pid: 88352) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-03 10:31:12,130] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 1 (pid: 489172) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-03 10:31:12,136] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-168.ec2.internal_88344_0' has failed to shutdown the rendezvous '12346536' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-03_10:31:05 + host : ip-26-0-171-168.ec2.internal + rank : 44 (local_rank: 4) + exitcode : 1 (pid: 88354) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-03_10:31:05 + host : ip-26-0-171-168.ec2.internal + rank : 45 (local_rank: 5) + exitcode : 1 (pid: 88355) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-03_10:31:05 + host : ip-26-0-171-168.ec2.internal + rank : 42 (local_rank: 2) + exitcode : 1 (pid: 88352) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-03 10:31:12,167] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-170-160.ec2.internal_489165_0' has failed to shutdown the rendezvous '12346536' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-03_10:31:05 + host : ip-26-0-170-160.ec2.internal + rank : 34 (local_rank: 2) + exitcode : 1 (pid: 489173) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-03_10:31:05 + host : ip-26-0-170-160.ec2.internal + rank : 36 (local_rank: 4) + exitcode : 1 (pid: 489175) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-03_10:31:05 + host : ip-26-0-170-160.ec2.internal + rank : 33 (local_rank: 1) + exitcode : 1 (pid: 489172) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-170-160: task 4: Exited with exit code 1 +srun: error: ip-26-0-170-143: task 3: Exited with exit code 1 +srun: error: ip-26-0-171-168: task 5: Exited with exit code 1 +[2024-12-03 10:31:12,777] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 844870) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-03 10:31:12,824] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-116.ec2.internal_844863_0' has failed to shutdown the rendezvous '12346536' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-03_10:31:05 + host : ip-26-0-172-116.ec2.internal + rank : 52 (local_rank: 4) + exitcode : 1 (pid: 844874) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-03_10:31:05 + host : ip-26-0-172-116.ec2.internal + rank : 48 (local_rank: 0) + exitcode : 1 (pid: 844870) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-172-116: task 7: Exited with exit code 1 +[2024-12-03 10:31:13,568] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 388544) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-03 10:31:13,615] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-73.ec2.internal_388537_0' has failed to shutdown the rendezvous '12346536' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-03_10:31:10 + host : ip-26-0-172-73.ec2.internal + rank : 57 (local_rank: 1) + exitcode : 1 (pid: 388545) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-03_10:31:10 + host : ip-26-0-172-73.ec2.internal + rank : 58 (local_rank: 2) + exitcode : 1 (pid: 388546) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-03_10:31:10 + host : ip-26-0-172-73.ec2.internal + rank : 59 (local_rank: 3) + exitcode : 1 (pid: 388547) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-03_10:31:10 + host : ip-26-0-172-73.ec2.internal + rank : 60 (local_rank: 4) + exitcode : 1 (pid: 388548) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-03_10:31:10 + host : ip-26-0-172-73.ec2.internal + rank : 61 (local_rank: 5) + exitcode : 1 (pid: 388549) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-03_10:31:10 + host : ip-26-0-172-73.ec2.internal + rank : 63 (local_rank: 7) + exitcode : 1 (pid: 388551) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-03_10:31:10 + host : ip-26-0-172-73.ec2.internal + rank : 56 (local_rank: 0) + exitcode : 1 (pid: 388544) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-172-73: task 6: Exited with exit code 1 diff --git a/logs/13132022-smolm2-bench.out b/logs/13132022-smolm2-bench.out new file mode 100644 index 0000000000000000000000000000000000000000..2bdf1067d887489549b051c7d6c849de1ddfa2bd --- /dev/null +++ b/logs/13132022-smolm2-bench.out @@ -0,0 +1,742 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-166-[15,36]' ++ export 'NODELIST=ip-26-0-166-15 +ip-26-0-166-36' ++ NODELIST='ip-26-0-166-15 +ip-26-0-166-36' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-166-[15,36]' ++ export MASTER_NODE=ip-26-0-166-15 ++ MASTER_NODE=ip-26-0-166-15 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-166-15' +Master node: ip-26-0-166-15 ++ echo 'All nodes: ip-26-0-166-15 +ip-26-0-166-36' +All nodes: ip-26-0-166-15 +ip-26-0-166-36 ++ echo 'World size: 16' +World size: 16 ++ srun torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13132022 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-166-15:12356 run_train.py --config-file examples/config_ali.yaml +[2024-12-07 23:09:50,613] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-07 23:09:50,625] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-07 23:09:50,613] torch.distributed.run: [WARNING] +[2024-12-07 23:09:50,613] torch.distributed.run: [WARNING] ***************************************** +[2024-12-07 23:09:50,613] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-07 23:09:50,613] torch.distributed.run: [WARNING] ***************************************** +[2024-12-07 23:09:50,625] torch.distributed.run: [WARNING] +[2024-12-07 23:09:50,625] torch.distributed.run: [WARNING] ***************************************** +[2024-12-07 23:09:50,625] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-07 23:09:50,625] torch.distributed.run: [WARNING] ***************************************** +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Config: +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Config(general=GeneralArgs(project='ali-darija', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: run='1p46G-gemma-fp-dedup-rehydr-ar-350BT-seed-6', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: seed=6, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: step=134000, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: consumed_train_samples=171520000, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: benchmark_csv_path=None, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: ignore_sanity_checks=True), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: parallelism=ParallelismArgs(dp=16, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pp=1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tp=1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pp_engine=, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tp_mode=, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tp_linear_async_communication=True, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: recompute_layer=False, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tp_recompute_allgather=True, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: expert_parallel_size=1), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: eos_token_id=2, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: hidden_act='silu', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: hidden_size=2048, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: initializer_range=0.02, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: intermediate_size=8192, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: is_llama_config=True, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: max_position_embeddings=2048, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_attention_heads=32, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_hidden_layers=14, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_key_value_heads=32, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pad_token_id=None, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pretraining_tp=1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rms_norm_eps=1e-05, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_scaling=None, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_theta=10000.0, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_interleaved=False, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tie_word_embeddings=True, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: use_cache=True, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: vocab_size=256008), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: init_method=RandomInit(std=0.02), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: dtype=torch.bfloat16, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: make_vocab_size_divisible_by=1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: ddp_bucket_cap_mb=25), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tokenizer=TokenizerArgs(tokenizer_name_or_path='google/gemma-7b', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tokenizer_revision=None, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tokenizer_max_length=None), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: checkpoint_interval=1000, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: save_initial_state=False, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: save_final_state=False, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: resume_checkpoint_path=xPath('checkpoints/1p46G-gemma-fp-dedup-rehydr-ar-350BT-seed-6/107000'), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: checkpoints_path_is_shared_file_system=False), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: logging=LoggingArgs(log_level='info', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: log_level_replica='info', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: iteration_step_info_interval=1), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tokens=TokensArgs(sequence_length=2048, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: train_steps=108734, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: micro_batch_size=8, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: batch_accumulation_per_replica=1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: val_check_interval=-1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: limit_val_batches=0, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: limit_test_batches=0), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: adam_beta1=0.9, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: adam_beta2=0.95, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: torch_adam_is_fused=True, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: name='adamW'), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: zero_stage=0, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: weight_decay=0.1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: clip_grad=1.0, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: accumulate_grad_in_fp32=True, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0007, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lr_warmup_steps=6700, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lr_warmup_style='linear', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lr_decay_style='linear', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lr_decay_steps=1733, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lr_decay_starting_step=107000, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: min_decay_lr=0)), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: start_training_step=1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: data=DataArgs(dataset=None, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: seed=42, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_loading_workers=1)), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: DatasetStageArgs(name='Finetuning Stage', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: start_training_step=107001, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: data=DataArgs(dataset=PretrainDatasetsArgs(hf_dataset_or_datasets='roneneldan/TinyStories', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: hf_dataset_splits='train', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: hf_dataset_config_name=None, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: dataset_processing_num_proc_per_process=1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: dataset_overwrite_cache=False, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: text_column_name='text'), +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: seed=42, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_loading_workers=1))], +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: profiler=None, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lighteval=None, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: s3_upload=None) +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Model Config: +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: LlamaConfig(bos_token_id=1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: eos_token_id=2, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: hidden_act='silu', +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: hidden_size=2048, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: initializer_range=0.02, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: intermediate_size=8192, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: is_llama_config=True, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: max_position_embeddings=2048, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_attention_heads=32, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_hidden_layers=14, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_key_value_heads=32, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pad_token_id=None, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pretraining_tp=1, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rms_norm_eps=1e-05, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_scaling=None, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_theta=10000.0, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_interleaved=False, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tie_word_embeddings=True, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: use_cache=True, +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: vocab_size=256008) +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Building model.. +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Initialize RoPE Theta = 10000.0 +12/07/2024 23:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Setting PP block ranks... +12/07/2024 23:10:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Total number of parameters: 1.46G (2792.14MiB) +12/07/2024 23:10:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Local number of parameters: 1.46G (2792.14MiB) +12/07/2024 23:10:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [After model building] Memory usage: 2792.17MiB. Peak allocated: 5504.00MiB Peak reserved: 22914.00MiB +12/07/2024 23:10:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Loading weights from checkpoints/1p46G-gemma-fp-dedup-rehydr-ar-350BT-seed-6/107000 + Loading weights: 0%| | 0/87 [00:00 + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +12/07/2024 23:10:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: CheckpointMetadata(version=, tp=1, dp=64, metas=TrainingMetadata(consumed_train_samples=136960000, last_train_step=107000, last_stage_idx=0, data_stages=[DataStageMetadata(name='Stable Training Stage', start_training_step=1, consumed_train_samples=136960000)]), custom_metas=None) +12/07/2024 23:10:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [Training Plan] Stage Stable Training Stage has 0 remaining training steps and has consumed 136960000 samples +12/07/2024 23:10:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Using dummy data generator +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 234, in + dataloader = get_dataloader(trainer) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 190, in get_dataloader + consumed_train_samples is not None +AssertionError: Cannot find consumed_train_samples for stage 107001 in the checkpoint +[2024-12-07 23:10:51,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447393 closing signal SIGTERM +[2024-12-07 23:10:51,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447394 closing signal SIGTERM +[2024-12-07 23:10:51,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447395 closing signal SIGTERM +[2024-12-07 23:10:51,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447396 closing signal SIGTERM +[2024-12-07 23:10:51,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447398 closing signal SIGTERM +[2024-12-07 23:10:51,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447400 closing signal SIGTERM +[2024-12-07 23:10:51,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1210346 closing signal SIGTERM +[2024-12-07 23:10:51,809] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1210348 closing signal SIGTERM +[2024-12-07 23:10:51,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1210349 closing signal SIGTERM +[2024-12-07 23:10:51,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1210350 closing signal SIGTERM +[2024-12-07 23:10:51,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1210351 closing signal SIGTERM +[2024-12-07 23:10:51,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1210352 closing signal SIGTERM +[2024-12-07 23:10:51,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1210353 closing signal SIGTERM +[2024-12-07 23:10:53,039] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 1 (pid: 1210347) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-07_23:10:51 + host : ip-26-0-166-36.ec2.internal + rank : 9 (local_rank: 1) + exitcode : 1 (pid: 1210347) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-07 23:10:53,177] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 4 (pid: 1447397) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-07_23:10:51 + host : ip-26-0-166-15.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 1447399) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-07_23:10:51 + host : ip-26-0-166-15.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 1447397) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-166-36: task 1: Exited with exit code 1 +srun: error: ip-26-0-166-15: task 0: Exited with exit code 1 diff --git a/logs/13264842-bench_1.34G_dp8_tp2_pp1_acc32_mbs2_seq8192_zero1_tpmodeRED_vocab131k.out b/logs/13264842-bench_1.34G_dp8_tp2_pp1_acc32_mbs2_seq8192_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..ba575a914abe76d6014c893d20a9f019cd7e70cf --- /dev/null +++ b/logs/13264842-bench_1.34G_dp8_tp2_pp1_acc32_mbs2_seq8192_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,634 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207,ip-26-0-170-132 ++ export 'NODELIST=ip-26-0-169-207 +ip-26-0-170-132' ++ NODELIST='ip-26-0-169-207 +ip-26-0-170-132' +++ scontrol show hostnames ip-26-0-169-207,ip-26-0-170-132 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207 +ip-26-0-170-132' +All nodes: ip-26-0-169-207 +ip-26-0-170-132 ++ echo 'World size: 16' +World size: 16 ++ srun torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13264842 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 run_train.py --config-file benchmark/configs/config_1.34G_dp8_tp2_pp1_acc32_mbs2_seq8192_zero1_tpmodeRED_vocab131k.yaml +[2024-12-09 23:26:44,439] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-09 23:26:44,439] torch.distributed.run: [WARNING] +[2024-12-09 23:26:44,439] torch.distributed.run: [WARNING] ***************************************** +[2024-12-09 23:26:44,439] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-09 23:26:44,439] torch.distributed.run: [WARNING] ***************************************** +[2024-12-09 23:26:44,512] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-09 23:26:44,512] torch.distributed.run: [WARNING] +[2024-12-09 23:26:44,512] torch.distributed.run: [WARNING] ***************************************** +[2024-12-09 23:26:44,512] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-09 23:26:44,512] torch.distributed.run: [WARNING] ***************************************** +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config: +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config(general=GeneralArgs(project='debug', +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: run='1.34G_dp8_tp2_pp1_acc32_mbs2_seq8192_zero1_tpmodeRED_vocab131k', +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: step=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: consumed_train_samples=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ignore_sanity_checks=True), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: parallelism=ParallelismArgs(dp=8, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp=1, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp=2, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp_engine=, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_mode=, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_linear_async_communication=True, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: recompute_layer=False, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_recompute_allgather=True, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: expert_parallel_size=1), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=8192, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=131072), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: init_method=RandomInit(std=0.02), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: dtype=torch.bfloat16, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: make_vocab_size_divisible_by=1, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ddp_bucket_cap_mb=25), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_revision=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_max_length=None), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoint_interval=10000, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_initial_state=False, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_final_state=False, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: resume_checkpoint_path=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints_path_is_shared_file_system=False), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: logging=LoggingArgs(log_level='info', +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: log_level_replica='info', +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration_step_info_interval=1), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokens=TokensArgs(sequence_length=8192, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: train_steps=100, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: micro_batch_size=2, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: batch_accumulation_per_replica=32, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: val_check_interval=100, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_val_batches=0, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_test_batches=0), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta1=0.9, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta2=0.95, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: torch_adam_is_fused=True, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: name='adamW'), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: zero_stage=1, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: weight_decay=0.01, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: clip_grad=1.0, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: accumulate_grad_in_fp32=True, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_steps=2, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_style='linear', +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_style='cosine', +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_steps=13, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_starting_step=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: min_decay_lr=1e-05)), +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: start_training_step=1, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data=DataArgs(dataset=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_loading_workers=1))], +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: profiler=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lighteval=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: s3_upload=None) +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Model Config: +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: LlamaConfig(bos_token_id=0, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=8192, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=131072) +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Building model.. +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Initialize RoPE Theta = 10000.0 +12/09/2024 23:27:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Setting PP block ranks... +12/09/2024 23:27:26 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: Local number of parameters: 621M (1184.13MiB) +12/09/2024 23:27:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Total number of parameters: 1.24G (2368.26MiB) +12/09/2024 23:27:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Local number of parameters: 621M (1184.13MiB) +12/09/2024 23:27:26 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: [After model building] Memory usage: 1184.15MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/09/2024 23:27:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [After model building] Memory usage: 1184.15MiB. Peak allocated: 5504.00MiB Peak reserved: 21890.00MiB +12/09/2024 23:27:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: No checkpoint path provided. +12/09/2024 23:27:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Parametrizing model parameters using StandardParametrizator +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Optimizer Building] Using LearningRateForSP as learning rate +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] Size of optimizer params per rank: +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 0 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 1 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 2 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 3 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 4 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 5 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 6 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 7 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Using dummy data generator +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] There are 1 training stages +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Stage Stable Training Stage] start from step 1 +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: +12/09/2024 23:27:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Start training] datetime: 2024-12-09 23:27:33.956282 | mbs: 2 | grad_accum: 32 | global_batch_size: 512 | sequence_length: 8192 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/09/2024 23:27:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/09/2024 23:27:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3848.44MiB. Peak allocated 5504.00MiB. Peak reserved: 24260.00MiB +12/09/2024 23:27:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4056.74MiB. Peak allocated 23850.64MiB. Peak reserved: 30198.00MiB +12/09/2024 23:27:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 9.44K | tokens_per_sec: 444K | tokens_per_sec_per_gpu: 27.8K | global_batch_size: 512 | lm_loss: 12.2 | lr: 0.00015 | model_tflops_per_gpu: 296 | hardware_tflops_per_gpu: 296 | grad_norm: 0.414 | cuda_memory_allocated: 4.89G | cuda_max_memory_reserved: 31.7G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.5G | hd_free_memory_tb: 244G +12/09/2024 23:27:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4660.10MiB. Peak allocated 7176.37MiB. Peak reserved: 30214.00MiB +12/09/2024 23:27:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4660.11MiB. Peak allocated 24454.01MiB. Peak reserved: 30230.00MiB +12/09/2024 23:27:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 7.52K | tokens_per_sec: 558K | tokens_per_sec_per_gpu: 34.9K | global_batch_size: 512 | lm_loss: 12.2 | lr: 0.0003 | model_tflops_per_gpu: 372 | hardware_tflops_per_gpu: 372 | grad_norm: 0.414 | cuda_memory_allocated: 4.89G | cuda_max_memory_reserved: 31.7G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.5G | hd_free_memory_tb: 244G +12/09/2024 23:27:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4660.10MiB. Peak allocated 7176.39MiB. Peak reserved: 30230.00MiB +12/09/2024 23:28:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4660.11MiB. Peak allocated 24454.01MiB. Peak reserved: 30230.00MiB +num_paramsnum_paramsnum_params + + +num_params{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576}{'total': 1241649152, 'local': 620824576} + + +num_params{'total': 1241649152, 'local': 620824576} + +num_paramsnum_params + +{'total': 1241649152, 'local': 620824576}{'total': 1241649152, 'local': 620824576} + +num_params{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +num_paramsnum_params + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} +num_params +num_params{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} +12/09/2024 23:28:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 3 / 100 | consumed_tokens: 12.6M | elapsed_time_per_iteration_ms: 7.53K | tokens_per_sec: 557K | tokens_per_sec_per_gpu: 34.8K | global_batch_size: 512 | lm_loss: 12.2 | lr: 0.000296 | model_tflops_per_gpu: 371 | hardware_tflops_per_gpu: 371 | grad_norm: 0.408 | cuda_memory_allocated: 4.89G | cuda_max_memory_reserved: 31.7G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.5G | hd_free_memory_tb: 244G +num_params +{'total': 1241649152, 'local': 620824576} +12/09/2024 23:28:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/09/2024 23:28:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: | -------- | -------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | --------------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/09/2024 23:28:06 [INFO|DP=5|PP=0|TP=1|ip-26-0-170-132]: Throughput logging complete +12/09/2024 23:28:06 [INFO|DP=4|PP=0|TP=1|ip-26-0-170-132]: Throughput logging complete +12/09/2024 23:28:06 [INFO|DP=7|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +12/09/2024 23:28:06 [INFO|DP=6|PP=0|TP=1|ip-26-0-170-132]: Throughput logging complete +12/09/2024 23:28:06 [INFO|DP=4|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +12/09/2024 23:28:06 [INFO|DP=6|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +12/09/2024 23:28:06 [INFO|DP=7|PP=0|TP=1|ip-26-0-170-132]: Throughput logging complete +12/09/2024 23:28:06 [INFO|DP=5|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +slurmstepd: error: *** JOB 13264842 ON ip-26-0-169-207 CANCELLED AT 2024-12-09T23:28:06 *** +[2024-12-09 23:28:06,919] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-09 23:28:06,920] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1758549 closing signal SIGTERM +[2024-12-09 23:28:06,920] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1758550 closing signal SIGTERM +[2024-12-09 23:28:06,920] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1758551 closing signal SIGTERM +[2024-12-09 23:28:06,920] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1758552 closing signal SIGTERM +[2024-12-09 23:28:06,923] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1758553 closing signal SIGTERM diff --git a/logs/13264900-bench_1.14G_dp8_tp2_pp1_acc2_mbs8_seq32768_zero1_tpmodeRED_vocab32k.out b/logs/13264900-bench_1.14G_dp8_tp2_pp1_acc2_mbs8_seq32768_zero1_tpmodeRED_vocab32k.out new file mode 100644 index 0000000000000000000000000000000000000000..6cf8661bb61a9f136e09e18705b5bd7ca808ca15 --- /dev/null +++ b/logs/13264900-bench_1.14G_dp8_tp2_pp1_acc2_mbs8_seq32768_zero1_tpmodeRED_vocab32k.out @@ -0,0 +1,1563 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-169-[86,132]' ++ export 'NODELIST=ip-26-0-169-86 +ip-26-0-169-132' ++ NODELIST='ip-26-0-169-86 +ip-26-0-169-132' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-169-[86,132]' ++ export MASTER_NODE=ip-26-0-169-86 ++ MASTER_NODE=ip-26-0-169-86 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-169-86' +Master node: ip-26-0-169-86 ++ echo 'All nodes: ip-26-0-169-86 +ip-26-0-169-132' +All nodes: ip-26-0-169-86 +ip-26-0-169-132 ++ echo 'World size: 16' +World size: 16 ++ srun torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13264900 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-86:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp8_tp2_pp1_acc2_mbs8_seq32768_zero1_tpmodeRED_vocab32k.yaml +[2024-12-09 23:37:27,191] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-09 23:37:27,191] torch.distributed.run: [WARNING] +[2024-12-09 23:37:27,191] torch.distributed.run: [WARNING] ***************************************** +[2024-12-09 23:37:27,191] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-09 23:37:27,191] torch.distributed.run: [WARNING] ***************************************** +[2024-12-09 23:37:27,288] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-09 23:37:27,288] torch.distributed.run: [WARNING] +[2024-12-09 23:37:27,288] torch.distributed.run: [WARNING] ***************************************** +[2024-12-09 23:37:27,288] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-09 23:37:27,288] torch.distributed.run: [WARNING] ***************************************** +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Config: +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Config(general=GeneralArgs(project='debug', +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: run='1.14G_dp8_tp2_pp1_acc2_mbs8_seq32768_zero1_tpmodeRED_vocab32k', +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: seed=42, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: step=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: consumed_train_samples=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: ignore_sanity_checks=True), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: parallelism=ParallelismArgs(dp=8, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pp=1, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp=2, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pp_engine=, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp_mode=, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp_linear_async_communication=True, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: recompute_layer=False, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp_recompute_allgather=True, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: expert_parallel_size=1), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: eos_token_id=0, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_act='silu', +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_size=2048, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: initializer_range=0.02, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: intermediate_size=8192, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: is_llama_config=True, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: max_position_embeddings=32768, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_attention_heads=32, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_hidden_layers=16, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_key_value_heads=8, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pad_token_id=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pretraining_tp=1, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rms_norm_eps=1e-05, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_scaling=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_theta=10000.0, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_interleaved=False, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tie_word_embeddings=True, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: use_cache=True, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: vocab_size=32768), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: init_method=RandomInit(std=0.02), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: dtype=torch.bfloat16, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: make_vocab_size_divisible_by=1, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: ddp_bucket_cap_mb=25), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokenizer_revision=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokenizer_max_length=None), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: checkpoint_interval=10000, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: save_initial_state=False, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: save_final_state=False, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: resume_checkpoint_path=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: checkpoints_path_is_shared_file_system=False), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: logging=LoggingArgs(log_level='info', +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: log_level_replica='info', +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: iteration_step_info_interval=1), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokens=TokensArgs(sequence_length=32768, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: train_steps=100, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: micro_batch_size=8, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: batch_accumulation_per_replica=2, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: val_check_interval=100, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: limit_val_batches=0, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: limit_test_batches=0), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: adam_beta1=0.9, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: adam_beta2=0.95, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: torch_adam_is_fused=True, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: name='adamW'), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: zero_stage=1, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: weight_decay=0.01, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: clip_grad=1.0, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: accumulate_grad_in_fp32=True, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_warmup_steps=2, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_warmup_style='linear', +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_decay_style='cosine', +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_decay_steps=13, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_decay_starting_step=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: min_decay_lr=1e-05)), +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: start_training_step=1, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: data=DataArgs(dataset=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: seed=42, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_loading_workers=1))], +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: profiler=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lighteval=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: s3_upload=None) +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Model Config: +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: LlamaConfig(bos_token_id=0, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: eos_token_id=0, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_act='silu', +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_size=2048, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: initializer_range=0.02, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: intermediate_size=8192, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: is_llama_config=True, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: max_position_embeddings=32768, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_attention_heads=32, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_hidden_layers=16, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_key_value_heads=8, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pad_token_id=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pretraining_tp=1, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rms_norm_eps=1e-05, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_scaling=None, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_theta=10000.0, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_interleaved=False, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tie_word_embeddings=True, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: use_cache=True, +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: vocab_size=32768) +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Building model.. +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Initialize RoPE Theta = 10000.0 +12/09/2024 23:38:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Setting PP block ranks... +12/09/2024 23:38:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Total number of parameters: 1.04G (1984.26MiB) +12/09/2024 23:38:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Local number of parameters: 520M (992.13MiB) +12/09/2024 23:38:11 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-132]: Local number of parameters: 520M (992.13MiB) +12/09/2024 23:38:11 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-132]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/09/2024 23:38:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/09/2024 23:38:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: No checkpoint path provided. +12/09/2024 23:38:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Parametrizing model parameters using StandardParametrizator +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [Optimizer Building] Using LearningRateForSP as learning rate +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] Size of optimizer params per rank: +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] DP Rank 0 has 65M out of 520M (12.50%) params' optimizer states +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] DP Rank 1 has 65M out of 520M (12.50%) params' optimizer states +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] DP Rank 2 has 65M out of 520M (12.50%) params' optimizer states +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] DP Rank 3 has 65M out of 520M (12.50%) params' optimizer states +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] DP Rank 4 has 65M out of 520M (12.50%) params' optimizer states +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] DP Rank 5 has 65M out of 520M (12.50%) params' optimizer states +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] DP Rank 6 has 65M out of 520M (12.50%) params' optimizer states +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] DP Rank 7 has 65M out of 520M (12.50%) params' optimizer states +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Using dummy data generator +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [Training Plan] There are 1 training stages +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [Stage Stable Training Stage] start from step 1 +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: +12/09/2024 23:38:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [Start training] datetime: 2024-12-09 23:38:18.795321 | mbs: 8 | grad_accum: 2 | global_batch_size: 128 | sequence_length: 32768 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 3.98 GiB is free. Including non-PyTorch memory, this process has 75.34 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 524.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 3.04 GiB is free. Including non-PyTorch memory, this process has 76.28 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 1.26 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 3.78 GiB is free. Including non-PyTorch memory, this process has 75.54 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 524.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 2.76 GiB is free. Including non-PyTorch memory, this process has 76.56 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 1.51 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 3.99 GiB is free. Including non-PyTorch memory, this process has 75.33 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 268.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 3.75 GiB is free. Including non-PyTorch memory, this process has 75.57 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 524.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 3.73 GiB is free. Including non-PyTorch memory, this process has 75.59 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 780.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + hidden_encoder_states = encoder_block(**hidden_encoder_states) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + sharded_logits = self.model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + Traceback (most recent call last): +hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return func(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 3.78 GiB is free. Including non-PyTorch memory, this process has 75.54 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 524.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return self._call_impl(*args, **kwargs) + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + output = self.pp_block(**new_kwargs)return super().apply(*args, **kwargs) # type: ignore[misc] + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + gathered_output = torch.empty( + return func(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 3.26 GiB is free. Including non-PyTorch memory, this process has 76.06 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 1.01 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 3.49 GiB is free. Including non-PyTorch memory, this process has 75.83 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 780.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.act(gate_states) * up_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/activations.py", line 149, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return nn.functional.silu(input) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/functional.py", line 2072, in silu + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return torch._C._nn.silu(input) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 1.94 MiB is free. Including non-PyTorch memory, this process has 79.32 GiB memory in use. Of the allocated memory 68.31 GiB is allocated by PyTorch, and 268.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 3.89 GiB is free. Including non-PyTorch memory, this process has 75.43 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 268.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 3.39 GiB is free. Including non-PyTorch memory, this process has 75.93 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 780.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 3.76 GiB is free. Including non-PyTorch memory, this process has 75.56 GiB memory in use. Of the allocated memory 64.31 GiB is allocated by PyTorch, and 524.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/09/2024 23:38:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/09/2024 23:38:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Memory usage: 3224.44MiB. Peak allocated 5504.00MiB. Peak reserved: 37188.00MiB +[2024-12-09 23:38:28,385] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 567082 closing signal SIGTERM +[2024-12-09 23:38:28,386] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 567083 closing signal SIGTERM +[2024-12-09 23:38:28,386] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 341397) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-86.ec2.internal + rank : 9 (local_rank: 1) + exitcode : 1 (pid: 341398) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-86.ec2.internal + rank : 10 (local_rank: 2) + exitcode : 1 (pid: 341399) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-86.ec2.internal + rank : 11 (local_rank: 3) + exitcode : 1 (pid: 341400) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-86.ec2.internal + rank : 12 (local_rank: 4) + exitcode : 1 (pid: 341401) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-86.ec2.internal + rank : 13 (local_rank: 5) + exitcode : 1 (pid: 341402) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-86.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 341403) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-86.ec2.internal + rank : 15 (local_rank: 7) + exitcode : 1 (pid: 341404) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-86.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 341397) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-169-86: task 0: Exited with exit code 1 +[2024-12-09 23:38:31,818] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 2 (pid: 567084) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-09 23:38:31,867] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-132.ec2.internal_567075_0' has failed to shutdown the rendezvous '13264900' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-132.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 567085) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-132.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 567086) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-132.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 567087) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-132.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 567088) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-132.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 567089) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-09_23:38:28 + host : ip-26-0-169-132.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 567084) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-169-132: task 1: Exited with exit code 1 diff --git a/logs/13264903-bench_1.34G_dp8_tp2_pp1_acc2_mbs8_seq32768_zero1_tpmodeALL_vocab131k.out b/logs/13264903-bench_1.34G_dp8_tp2_pp1_acc2_mbs8_seq32768_zero1_tpmodeALL_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..fa199827634d8400f95d929ca0a940c997fc326c --- /dev/null +++ b/logs/13264903-bench_1.34G_dp8_tp2_pp1_acc2_mbs8_seq32768_zero1_tpmodeALL_vocab131k.out @@ -0,0 +1,1472 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-171-[168,230]' ++ export 'NODELIST=ip-26-0-171-168 +ip-26-0-171-230' ++ NODELIST='ip-26-0-171-168 +ip-26-0-171-230' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-171-[168,230]' ++ export MASTER_NODE=ip-26-0-171-168 ++ MASTER_NODE=ip-26-0-171-168 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-171-168' +Master node: ip-26-0-171-168 ++ echo 'All nodes: ip-26-0-171-168 +ip-26-0-171-230' +All nodes: ip-26-0-171-168 +ip-26-0-171-230 ++ echo 'World size: 16' +World size: 16 ++ srun torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13264903 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-171-168:12356 run_train.py --config-file benchmark/configs/config_1.34G_dp8_tp2_pp1_acc2_mbs8_seq32768_zero1_tpmodeALL_vocab131k.yaml +[2024-12-09 23:37:27,152] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-09 23:37:27,152] torch.distributed.run: [WARNING] +[2024-12-09 23:37:27,152] torch.distributed.run: [WARNING] ***************************************** +[2024-12-09 23:37:27,152] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-09 23:37:27,152] torch.distributed.run: [WARNING] ***************************************** +[2024-12-09 23:37:27,274] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-09 23:37:27,274] torch.distributed.run: [WARNING] +[2024-12-09 23:37:27,274] torch.distributed.run: [WARNING] ***************************************** +[2024-12-09 23:37:27,274] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-09 23:37:27,274] torch.distributed.run: [WARNING] ***************************************** +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Config: +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Config(general=GeneralArgs(project='debug', +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: run='1.34G_dp8_tp2_pp1_acc2_mbs8_seq32768_zero1_tpmodeALL_vocab131k', +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: seed=42, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: step=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: consumed_train_samples=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: ignore_sanity_checks=True), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: parallelism=ParallelismArgs(dp=8, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pp=1, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tp=2, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pp_engine=, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tp_mode=, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tp_linear_async_communication=False, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: recompute_layer=False, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tp_recompute_allgather=True, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: expert_parallel_size=1), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: eos_token_id=0, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: hidden_act='silu', +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: hidden_size=2048, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: initializer_range=0.02, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: intermediate_size=8192, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: is_llama_config=True, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: max_position_embeddings=32768, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_attention_heads=32, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_hidden_layers=16, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_key_value_heads=8, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pad_token_id=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pretraining_tp=1, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rms_norm_eps=1e-05, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_scaling=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_theta=10000.0, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_interleaved=False, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tie_word_embeddings=True, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: use_cache=True, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: vocab_size=131072), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: init_method=RandomInit(std=0.02), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: dtype=torch.bfloat16, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: make_vocab_size_divisible_by=1, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: ddp_bucket_cap_mb=25), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tokenizer_revision=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tokenizer_max_length=None), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: checkpoint_interval=10000, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: save_initial_state=False, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: save_final_state=False, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: resume_checkpoint_path=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: checkpoints_path_is_shared_file_system=False), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: logging=LoggingArgs(log_level='info', +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: log_level_replica='info', +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: iteration_step_info_interval=1), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tokens=TokensArgs(sequence_length=32768, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: train_steps=100, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: micro_batch_size=8, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: batch_accumulation_per_replica=2, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: val_check_interval=100, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: limit_val_batches=0, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: limit_test_batches=0), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: adam_beta1=0.9, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: adam_beta2=0.95, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: torch_adam_is_fused=True, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: name='adamW'), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: zero_stage=1, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: weight_decay=0.01, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: clip_grad=1.0, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: accumulate_grad_in_fp32=True, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lr_warmup_steps=2, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lr_warmup_style='linear', +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lr_decay_style='cosine', +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lr_decay_steps=13, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lr_decay_starting_step=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: min_decay_lr=1e-05)), +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: start_training_step=1, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: data=DataArgs(dataset=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: seed=42, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_loading_workers=1))], +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: profiler=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lighteval=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: s3_upload=None) +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Model Config: +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: LlamaConfig(bos_token_id=0, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: eos_token_id=0, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: hidden_act='silu', +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: hidden_size=2048, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: initializer_range=0.02, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: intermediate_size=8192, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: is_llama_config=True, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: max_position_embeddings=32768, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_attention_heads=32, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_hidden_layers=16, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_key_value_heads=8, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pad_token_id=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pretraining_tp=1, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rms_norm_eps=1e-05, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_scaling=None, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_theta=10000.0, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_interleaved=False, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tie_word_embeddings=True, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: use_cache=True, +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: vocab_size=131072) +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Building model.. +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Initialize RoPE Theta = 10000.0 +12/09/2024 23:38:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Setting PP block ranks... +12/09/2024 23:38:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Total number of parameters: 1.24G (2368.26MiB) +12/09/2024 23:38:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Local number of parameters: 621M (1184.13MiB) +12/09/2024 23:38:12 [INFO|DP=0|PP=0|TP=1|ip-26-0-171-168]: Local number of parameters: 621M (1184.13MiB) +12/09/2024 23:38:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [After model building] Memory usage: 1184.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/09/2024 23:38:12 [INFO|DP=0|PP=0|TP=1|ip-26-0-171-168]: [After model building] Memory usage: 1184.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/09/2024 23:38:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: No checkpoint path provided. +12/09/2024 23:38:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Parametrizing model parameters using StandardParametrizator +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [Optimizer Building] Using LearningRateForSP as learning rate +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [ZeRO sharding] Size of optimizer params per rank: +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [ZeRO sharding] DP Rank 0 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [ZeRO sharding] DP Rank 1 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [ZeRO sharding] DP Rank 2 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [ZeRO sharding] DP Rank 3 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [ZeRO sharding] DP Rank 4 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [ZeRO sharding] DP Rank 5 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [ZeRO sharding] DP Rank 6 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [ZeRO sharding] DP Rank 7 has 77.6M out of 621M (12.50%) params' optimizer states +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Using dummy data generator +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [Training Plan] There are 1 training stages +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [Stage Stable Training Stage] start from step 1 +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: +12/09/2024 23:38:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [Start training] datetime: 2024-12-09 23:38:19.748695 | mbs: 8 | grad_accum: 2 | global_batch_size: 128 | sequence_length: 32768 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 2.60 GiB is free. Including non-PyTorch memory, this process has 76.72 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 560.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 2.65 GiB is free. Including non-PyTorch memory, this process has 76.66 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 304.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 2.16 GiB is free. Including non-PyTorch memory, this process has 77.16 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 816.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 2.13 GiB is free. Including non-PyTorch memory, this process has 77.19 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 816.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step +return column_linear( +output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 1.51 GiB is free. Including non-PyTorch memory, this process has 77.80 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 1.30 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return forward_call(*args, **kwargs)hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 2.38 GiB is free. Including non-PyTorch memory, this process has 76.94 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 560.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 2.37 GiB is free. Including non-PyTorch memory, this process has 76.95 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 560.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 2.36 GiB is free. Including non-PyTorch memory, this process has 76.96 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 560.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return self._call_impl(*args, **kwargs) + trainer.train(dataloader) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 2.60 GiB is free. Including non-PyTorch memory, this process has 76.72 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 560.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 2.65 GiB is free. Including non-PyTorch memory, this process has 76.66 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 304.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 2.61 GiB is free. Including non-PyTorch memory, this process has 76.71 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 304.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 2.62 GiB is free. Including non-PyTorch memory, this process has 76.70 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 304.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 2.51 GiB is free. Including non-PyTorch memory, this process has 76.80 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 304.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 2.63 GiB is free. Including non-PyTorch memory, this process has 76.69 GiB memory in use. Of the allocated memory 65.65 GiB is allocated by PyTorch, and 304.80 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/09/2024 23:38:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/09/2024 23:38:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Memory usage: 3848.44MiB. Peak allocated 5504.00MiB. Peak reserved: 37572.00MiB +[2024-12-09 23:38:28,640] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2079616 closing signal SIGTERM +[2024-12-09 23:38:28,641] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2079617 closing signal SIGTERM +[2024-12-09 23:38:28,648] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2946437 closing signal SIGTERM +[2024-12-09 23:38:28,963] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2946436) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-230.ec2.internal + rank : 10 (local_rank: 2) + exitcode : 1 (pid: 2946438) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-230.ec2.internal + rank : 11 (local_rank: 3) + exitcode : 1 (pid: 2946439) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-230.ec2.internal + rank : 12 (local_rank: 4) + exitcode : 1 (pid: 2946440) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-230.ec2.internal + rank : 13 (local_rank: 5) + exitcode : 1 (pid: 2946441) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-230.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 2946442) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-230.ec2.internal + rank : 15 (local_rank: 7) + exitcode : 1 (pid: 2946443) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-230.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 2946436) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-171-230: task 1: Exited with exit code 1 +[2024-12-09 23:38:32,124] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 2 (pid: 2079618) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-168.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 2079619) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-168.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 2079620) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-168.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 2079621) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-168.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 2079622) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-168.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 2079623) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-09_23:38:28 + host : ip-26-0-171-168.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 2079618) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-171-168: task 0: Exited with exit code 1 diff --git a/logs/13264947-bench_1.34G_dp16_tp2_pp1_acc8_mbs16_seq2048_zero1_tpmodeALL_vocab131k.out b/logs/13264947-bench_1.34G_dp16_tp2_pp1_acc8_mbs16_seq2048_zero1_tpmodeALL_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..cb19ede6f6286754f137d461bc8eb681b496ea8d --- /dev/null +++ b/logs/13264947-bench_1.34G_dp16_tp2_pp1_acc8_mbs16_seq2048_zero1_tpmodeALL_vocab131k.out @@ -0,0 +1,748 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-163-58,ip-26-0-166-244,ip-26-0-167-[9,51]' ++ export 'NODELIST=ip-26-0-163-58 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51' ++ NODELIST='ip-26-0-163-58 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-163-58,ip-26-0-166-244,ip-26-0-167-[9,51]' ++ export MASTER_NODE=ip-26-0-163-58 ++ MASTER_NODE=ip-26-0-163-58 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-163-58' +Master node: ip-26-0-163-58 ++ echo 'All nodes: ip-26-0-163-58 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51' +All nodes: ip-26-0-163-58 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 ++ echo 'World size: 32' +World size: 32 ++ srun torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13264947 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-163-58:12356 run_train.py --config-file benchmark/configs/config_1.34G_dp16_tp2_pp1_acc8_mbs16_seq2048_zero1_tpmodeALL_vocab131k.yaml +[2024-12-10 00:23:59,640] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 00:23:59,649] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 00:23:59,652] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 00:23:59,641] torch.distributed.run: [WARNING] +[2024-12-10 00:23:59,641] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 00:23:59,641] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 00:23:59,641] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 00:23:59,652] torch.distributed.run: [WARNING] +[2024-12-10 00:23:59,652] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 00:23:59,652] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 00:23:59,652] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 00:23:59,649] torch.distributed.run: [WARNING] +[2024-12-10 00:23:59,649] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 00:23:59,649] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 00:23:59,649] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 00:23:59,703] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 00:23:59,703] torch.distributed.run: [WARNING] +[2024-12-10 00:23:59,703] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 00:23:59,703] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 00:23:59,703] torch.distributed.run: [WARNING] ***************************************** +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Config: +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Config(general=GeneralArgs(project='debug', +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: run='1.34G_dp16_tp2_pp1_acc8_mbs16_seq2048_zero1_tpmodeALL_vocab131k', +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: seed=42, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: step=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: consumed_train_samples=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: ignore_sanity_checks=True), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: parallelism=ParallelismArgs(dp=16, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: pp=1, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: tp=2, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: pp_engine=, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: tp_mode=, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: tp_linear_async_communication=False, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: recompute_layer=False, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: tp_recompute_allgather=True, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: expert_parallel_size=1), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: eos_token_id=0, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: hidden_act='silu', +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: hidden_size=2048, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: initializer_range=0.02, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: intermediate_size=8192, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: is_llama_config=True, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: max_position_embeddings=2048, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: num_attention_heads=32, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: num_hidden_layers=16, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: num_key_value_heads=8, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: pad_token_id=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: pretraining_tp=1, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: rms_norm_eps=1e-05, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: rope_scaling=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: rope_theta=10000.0, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: rope_interleaved=False, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: tie_word_embeddings=True, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: use_cache=True, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: vocab_size=131072), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: init_method=RandomInit(std=0.02), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: dtype=torch.bfloat16, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: make_vocab_size_divisible_by=1, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: ddp_bucket_cap_mb=25), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: tokenizer_revision=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: tokenizer_max_length=None), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: checkpoint_interval=10000, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: save_initial_state=False, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: save_final_state=False, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: resume_checkpoint_path=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: checkpoints_path_is_shared_file_system=False), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: logging=LoggingArgs(log_level='info', +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: log_level_replica='info', +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: iteration_step_info_interval=1), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: tokens=TokensArgs(sequence_length=2048, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: train_steps=100, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: micro_batch_size=16, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: batch_accumulation_per_replica=8, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: val_check_interval=100, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: limit_val_batches=0, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: limit_test_batches=0), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: adam_beta1=0.9, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: adam_beta2=0.95, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: torch_adam_is_fused=True, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: name='adamW'), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: zero_stage=1, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: weight_decay=0.01, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: clip_grad=1.0, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: accumulate_grad_in_fp32=True, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: lr_warmup_steps=2, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: lr_warmup_style='linear', +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: lr_decay_style='cosine', +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: lr_decay_steps=13, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: lr_decay_starting_step=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: min_decay_lr=1e-05)), +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: start_training_step=1, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: data=DataArgs(dataset=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: seed=42, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: num_loading_workers=1))], +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: profiler=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: lighteval=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: s3_upload=None) +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Model Config: +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: LlamaConfig(bos_token_id=0, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: eos_token_id=0, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: hidden_act='silu', +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: hidden_size=2048, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: initializer_range=0.02, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: intermediate_size=8192, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: is_llama_config=True, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: max_position_embeddings=2048, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: num_attention_heads=32, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: num_hidden_layers=16, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: num_key_value_heads=8, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: pad_token_id=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: pretraining_tp=1, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: rms_norm_eps=1e-05, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: rope_scaling=None, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: rope_theta=10000.0, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: rope_interleaved=False, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: tie_word_embeddings=True, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: use_cache=True, +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: vocab_size=131072) +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Building model.. +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Initialize RoPE Theta = 10000.0 +12/10/2024 00:24:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Setting PP block ranks... +12/10/2024 00:24:44 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Total number of parameters: 1.24G (2368.26MiB) +12/10/2024 00:24:44 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Local number of parameters: 621M (1184.13MiB) +12/10/2024 00:24:44 [INFO|DP=0|PP=0|TP=1|ip-26-0-163-58]: Local number of parameters: 621M (1184.13MiB) +12/10/2024 00:24:44 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [After model building] Memory usage: 1184.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/10/2024 00:24:44 [INFO|DP=0|PP=0|TP=1|ip-26-0-163-58]: [After model building] Memory usage: 1184.15MiB. Peak allocated: 5440.00MiB Peak reserved: 23874.00MiB +12/10/2024 00:24:44 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: No checkpoint path provided. +12/10/2024 00:24:44 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Parametrizing model parameters using StandardParametrizator +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [Optimizer Building] Using LearningRateForSP as learning rate +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] Size of optimizer params per rank: +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 0 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 1 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 2 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 3 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 4 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 5 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 6 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 7 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 8 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 9 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 10 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 11 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 12 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 13 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 14 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [ZeRO sharding] DP Rank 15 has 38.8M out of 621M (6.25%) params' optimizer states +12/10/2024 00:24:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/10/2024 00:24:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Using dummy data generator +12/10/2024 00:24:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [Training Plan] There are 1 training stages +12/10/2024 00:24:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [Stage Stable Training Stage] start from step 1 +12/10/2024 00:24:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: +12/10/2024 00:24:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: [Start training] datetime: 2024-12-10 00:24:52.470832 | mbs: 16 | grad_accum: 8 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/10/2024 00:25:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/10/2024 00:25:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Memory usage: 3700.43MiB. Peak allocated 5440.00MiB. Peak reserved: 25220.00MiB +12/10/2024 00:25:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Memory usage: 3768.99MiB. Peak allocated 48606.83MiB. Peak reserved: 61284.00MiB +12/10/2024 00:25:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 5.09K | tokens_per_sec: 824K | tokens_per_sec_per_gpu: 25.8K | global_batch_size: 2.05K | lm_loss: 12.2 | lr: 0.00015 | model_tflops_per_gpu: 213 | hardware_tflops_per_gpu: 213 | grad_norm: 0.431 | cuda_memory_allocated: 4.26G | cuda_max_memory_reserved: 64.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.7G | hd_free_memory_tb: 241G +12/10/2024 00:25:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Memory usage: 4065.07MiB. Peak allocated 6507.34MiB. Peak reserved: 61324.00MiB +12/10/2024 00:25:09 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Memory usage: 4065.08MiB. Peak allocated 48902.91MiB. Peak reserved: 61324.00MiB +12/10/2024 00:25:09 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 2.99K | tokens_per_sec: 1.4M | tokens_per_sec_per_gpu: 43.8K | global_batch_size: 2.05K | lm_loss: 12.2 | lr: 0.0003 | model_tflops_per_gpu: 362 | hardware_tflops_per_gpu: 362 | grad_norm: 0.43 | cuda_memory_allocated: 4.26G | cuda_max_memory_reserved: 64.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.7G | hd_free_memory_tb: 241G +12/10/2024 00:25:09 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Memory usage: 4065.07MiB. Peak allocated 6507.35MiB. Peak reserved: 61324.00MiB +12/10/2024 00:25:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Memory usage: 4065.08MiB. Peak allocated 48902.91MiB. Peak reserved: 61324.00MiB +num_paramsnum_params + +{'total': 1241649152, 'local': 620824576}{'total': 1241649152, 'local': 620824576} + +num_paramsnum_params + +num_params{'total': 1241649152, 'local': 620824576}{'total': 1241649152, 'local': 620824576} + + +num_params +num_params{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576}num_params + +num_params{'total': 1241649152, 'local': 620824576} + +num_params +{'total': 1241649152, 'local': 620824576} +num_paramsnum_paramsnum_params +num_params + +num_paramsnum_params{'total': 1241649152, 'local': 620824576} + + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +num_paramsnum_params + +num_params{'total': 1241649152, 'local': 620824576}{'total': 1241649152, 'local': 620824576} + + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576}num_params{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +num_params +num_params +{'total': 1241649152, 'local': 620824576} +num_params +num_params{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576}num_params + +{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} +num_params +num_params{'total': 1241649152, 'local': 620824576} + +num_params{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} +12/10/2024 00:25:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: iteration: 3 / 100 | consumed_tokens: 12.6M | elapsed_time_per_iteration_ms: 3K | tokens_per_sec: 1.4M | tokens_per_sec_per_gpu: 43.7K | global_batch_size: 2.05K | lm_loss: 12.2 | lr: 0.000296 | model_tflops_per_gpu: 361 | hardware_tflops_per_gpu: 361 | grad_norm: 0.425 | cuda_memory_allocated: 4.26G | cuda_max_memory_reserved: 64.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.7G | hd_free_memory_tb: 241G +num_params +{'total': 1241649152, 'local': 620824576} +12/10/2024 00:25:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/10/2024 00:25:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: | -------- | --------------------------------------------------------------- | ----- | ------- | --- | ----------- | ---- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | ----------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/10/2024 00:25:12 [INFO|DP=10|PP=0|TP=1|ip-26-0-167-51]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=4|PP=0|TP=1|ip-26-0-166-244]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=5|PP=0|TP=1|ip-26-0-166-244]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=7|PP=0|TP=0|ip-26-0-166-244]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=4|PP=0|TP=0|ip-26-0-166-244]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=15|PP=0|TP=0|ip-26-0-167-9]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=9|PP=0|TP=0|ip-26-0-167-51]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=8|PP=0|TP=1|ip-26-0-167-51]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=11|PP=0|TP=0|ip-26-0-167-51]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=8|PP=0|TP=0|ip-26-0-167-51]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=11|PP=0|TP=1|ip-26-0-167-51]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=6|PP=0|TP=1|ip-26-0-166-244]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=6|PP=0|TP=0|ip-26-0-166-244]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=10|PP=0|TP=0|ip-26-0-167-51]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=9|PP=0|TP=1|ip-26-0-167-51]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=7|PP=0|TP=1|ip-26-0-166-244]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=5|PP=0|TP=0|ip-26-0-166-244]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=12|PP=0|TP=0|ip-26-0-167-9]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=14|PP=0|TP=1|ip-26-0-167-9]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=12|PP=0|TP=1|ip-26-0-167-9]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=15|PP=0|TP=1|ip-26-0-167-9]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=14|PP=0|TP=0|ip-26-0-167-9]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=13|PP=0|TP=0|ip-26-0-167-9]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=13|PP=0|TP=1|ip-26-0-167-9]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: | 13264947 | 1.34G_dp16_tp2_pp1_acc8_mbs16_seq2048_zero1_tpmodeALL_vocab131k | 4 | 2048 | 16 | 8 | 2048 | 360.82 | 360.82 | 43712.12 | 247.19 | 212.45 | 190.66 | 459.85 | 265.20 | 265.41 | 6.35 | 59.89 | 16 | 1 | 2 | | TensorParallelLinearMode.ALL_REDUCE | False | 2048 | silu | 16 | 32 | 8 | 2048 | 131072 | True | torch.bfloat16 | 1 | 25 | True | 1.24G | 621M | +12/10/2024 00:25:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-58]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=0|PP=0|TP=1|ip-26-0-163-58]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=1|PP=0|TP=0|ip-26-0-163-58]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=3|PP=0|TP=1|ip-26-0-163-58]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=1|PP=0|TP=1|ip-26-0-163-58]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=3|PP=0|TP=0|ip-26-0-163-58]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=2|PP=0|TP=1|ip-26-0-163-58]: Throughput logging complete +12/10/2024 00:25:12 [INFO|DP=2|PP=0|TP=0|ip-26-0-163-58]: Throughput logging complete +slurmstepd: error: *** STEP 13264947.0 ON ip-26-0-163-58 CANCELLED AT 2024-12-10T00:25:13 *** +[2024-12-10 00:25:13,108] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 00:25:13,109] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81674 closing signal SIGTERM +[2024-12-10 00:25:13,109] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81675 closing signal SIGTERM +[2024-12-10 00:25:13,109] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81676 closing signal SIGTERM +[2024-12-10 00:25:13,108] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 00:25:13,109] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 00:25:13,110] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918914 closing signal SIGTERM +[2024-12-10 00:25:13,108] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2354328 closing signal SIGTERM +[2024-12-10 00:25:13,110] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918915 closing signal SIGTERM +[2024-12-10 00:25:13,110] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918916 closing signal SIGTERM +[2024-12-10 00:25:13,108] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2354329 closing signal SIGTERM +[2024-12-10 00:25:13,110] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918917 closing signal SIGTERM +[2024-12-10 00:25:13,109] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2354330 closing signal SIGTERM +slurmstepd: error: *** JOB 13264947 ON ip-26-0-163-58 CANCELLED AT 2024-12-10T00:25:13 *** +[2024-12-10 00:25:13,109] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81677 closing signal SIGTERM +[2024-12-10 00:25:13,110] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918918 closing signal SIGTERM +[2024-12-10 00:25:13,110] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918919 closing signal SIGTERM +[2024-12-10 00:25:13,109] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2354331 closing signal SIGTERM +[2024-12-10 00:25:13,110] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2354332 closing signal SIGTERM +[2024-12-10 00:25:13,109] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81678 closing signal SIGTERM +[2024-12-10 00:25:13,111] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2354333 closing signal SIGTERM +[2024-12-10 00:25:13,111] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2354334 closing signal SIGTERM +[2024-12-10 00:25:13,111] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81679 closing signal SIGTERM +[2024-12-10 00:25:13,111] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81680 closing signal SIGTERM +[2024-12-10 00:25:13,111] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81681 closing signal SIGTERM +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-10 00:25:13,111] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2354335 closing signal SIGTERM +[2024-12-10 00:25:13,113] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918920 closing signal SIGTERM +[2024-12-10 00:25:13,113] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918921 closing signal SIGTERM +[2024-12-10 00:25:13,118] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 00:25:13,118] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 120314 closing signal SIGTERM +[2024-12-10 00:25:13,118] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 120315 closing signal SIGTERM +[2024-12-10 00:25:13,118] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 120316 closing signal SIGTERM +[2024-12-10 00:25:13,119] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 120317 closing signal SIGTERM +[2024-12-10 00:25:13,120] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 120318 closing signal SIGTERM +[2024-12-10 00:25:13,120] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 120319 closing signal SIGTERM +[2024-12-10 00:25:13,120] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 120320 closing signal SIGTERM +[2024-12-10 00:25:13,122] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 120321 closing signal SIGTERM diff --git a/logs/13265039-bench_1.34G_dp32_tp2_pp1_acc8_mbs2_seq2048_zero1_tpmodeALL_vocab131k.out b/logs/13265039-bench_1.34G_dp32_tp2_pp1_acc8_mbs2_seq2048_zero1_tpmodeALL_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..c6ea692d22487d9a8447e07abd60f7989613635f --- /dev/null +++ b/logs/13265039-bench_1.34G_dp32_tp2_pp1_acc8_mbs2_seq2048_zero1_tpmodeALL_vocab131k.out @@ -0,0 +1,921 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-160-225,ip-26-0-166-[15,36,125,214,244],ip-26-0-167-[9,51]' ++ export 'NODELIST=ip-26-0-160-225 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51' ++ NODELIST='ip-26-0-160-225 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-160-225,ip-26-0-166-[15,36,125,214,244],ip-26-0-167-[9,51]' ++ export MASTER_NODE=ip-26-0-160-225 ++ MASTER_NODE=ip-26-0-160-225 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=8 ++ NNODES=8 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=64 ++ WORLD_SIZE=64 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-225' +Master node: ip-26-0-160-225 ++ echo 'All nodes: ip-26-0-160-225 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51' +All nodes: ip-26-0-160-225 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 ++ echo 'World size: 64' +World size: 64 ++ srun torchrun --nnodes=8 --nproc_per_node=8 --rdzv_id=13265039 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-225:12356 run_train.py --config-file benchmark/configs/config_1.34G_dp32_tp2_pp1_acc8_mbs2_seq2048_zero1_tpmodeALL_vocab131k.yaml +[2024-12-10 01:03:27,553] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 01:03:27,553] torch.distributed.run: [WARNING] +[2024-12-10 01:03:27,553] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,553] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 01:03:27,553] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 01:03:27,555] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,573] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 01:03:27,574] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,555] torch.distributed.run: [WARNING] +[2024-12-10 01:03:27,555] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,555] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 01:03:27,555] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 01:03:27,559] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,574] torch.distributed.run: [WARNING] +[2024-12-10 01:03:27,574] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,574] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 01:03:27,574] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,573] torch.distributed.run: [WARNING] +[2024-12-10 01:03:27,573] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,573] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 01:03:27,573] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,666] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 01:03:27,666] torch.distributed.run: [WARNING] +[2024-12-10 01:03:27,666] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:03:27,666] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 01:03:27,666] torch.distributed.run: [WARNING] ***************************************** +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Config: +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Config(general=GeneralArgs(project='debug', +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: run='1.34G_dp32_tp2_pp1_acc8_mbs2_seq2048_zero1_tpmodeALL_vocab131k', +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: seed=42, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: step=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: consumed_train_samples=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: ignore_sanity_checks=True), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: parallelism=ParallelismArgs(dp=32, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pp=1, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp=2, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pp_engine=, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp_mode=, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp_linear_async_communication=False, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: recompute_layer=False, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp_recompute_allgather=True, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: expert_parallel_size=1), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: eos_token_id=0, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_act='silu', +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_size=2048, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: initializer_range=0.02, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: intermediate_size=8192, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: is_llama_config=True, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: max_position_embeddings=2048, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_attention_heads=32, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_hidden_layers=16, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_key_value_heads=8, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pad_token_id=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pretraining_tp=1, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rms_norm_eps=1e-05, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_scaling=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_theta=10000.0, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_interleaved=False, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tie_word_embeddings=True, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: use_cache=True, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: vocab_size=131072), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: init_method=RandomInit(std=0.02), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: dtype=torch.bfloat16, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: make_vocab_size_divisible_by=1, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: ddp_bucket_cap_mb=25), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer_revision=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer_max_length=None), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoint_interval=10000, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: save_initial_state=False, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: save_final_state=False, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: resume_checkpoint_path=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoints_path_is_shared_file_system=False), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: logging=LoggingArgs(log_level='info', +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: log_level_replica='info', +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: iteration_step_info_interval=1), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokens=TokensArgs(sequence_length=2048, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: train_steps=100, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: micro_batch_size=2, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: batch_accumulation_per_replica=8, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: val_check_interval=100, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: limit_val_batches=0, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: limit_test_batches=0), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: adam_beta1=0.9, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: adam_beta2=0.95, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: torch_adam_is_fused=True, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: name='adamW'), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: zero_stage=1, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: weight_decay=0.01, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: clip_grad=1.0, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: accumulate_grad_in_fp32=True, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_warmup_steps=2, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_warmup_style='linear', +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_style='cosine', +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_steps=13, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_starting_step=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: min_decay_lr=1e-05)), +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: start_training_step=1, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: data=DataArgs(dataset=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: seed=42, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_loading_workers=1))], +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: profiler=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lighteval=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: s3_upload=None) +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Model Config: +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: LlamaConfig(bos_token_id=0, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: eos_token_id=0, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_act='silu', +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_size=2048, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: initializer_range=0.02, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: intermediate_size=8192, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: is_llama_config=True, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: max_position_embeddings=2048, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_attention_heads=32, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_hidden_layers=16, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_key_value_heads=8, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pad_token_id=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pretraining_tp=1, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rms_norm_eps=1e-05, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_scaling=None, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_theta=10000.0, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_interleaved=False, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tie_word_embeddings=True, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: use_cache=True, +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: vocab_size=131072) +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Building model.. +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Initialize RoPE Theta = 10000.0 +12/10/2024 01:04:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Setting PP block ranks... +12/10/2024 01:04:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Total number of parameters: 1.24G (2368.26MiB) +12/10/2024 01:04:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Local number of parameters: 621M (1184.13MiB) +12/10/2024 01:04:14 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-225]: Local number of parameters: 621M (1184.13MiB) +12/10/2024 01:04:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [After model building] Memory usage: 1184.15MiB. Peak allocated: 5408.00MiB Peak reserved: 27938.00MiB +12/10/2024 01:04:14 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-225]: [After model building] Memory usage: 1184.15MiB. Peak allocated: 5408.00MiB Peak reserved: 25890.00MiB +12/10/2024 01:04:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: No checkpoint path provided. +12/10/2024 01:04:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Parametrizing model parameters using StandardParametrizator +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Optimizer Building] Using LearningRateForSP as learning rate +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] Size of optimizer params per rank: +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 0 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 1 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 2 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 3 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 4 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 5 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 6 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 7 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 8 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 9 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 10 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 11 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 12 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 13 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 14 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 15 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 16 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 17 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 18 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 19 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 20 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 21 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 22 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 23 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 24 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 25 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 26 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 27 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 28 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 29 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 30 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 31 has 19.4M out of 621M (3.12%) params' optimizer states +12/10/2024 01:04:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/10/2024 01:04:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Using dummy data generator +12/10/2024 01:04:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Training Plan] There are 1 training stages +12/10/2024 01:04:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Stage Stable Training Stage] start from step 1 +12/10/2024 01:04:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: +12/10/2024 01:04:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Start training] datetime: 2024-12-10 01:04:23.351592 | mbs: 2 | grad_accum: 8 | global_batch_size: 512 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/10/2024 01:04:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/10/2024 01:04:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3626.42MiB. Peak allocated 5408.00MiB. Peak reserved: 30308.00MiB +12/10/2024 01:04:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3694.49MiB. Peak allocated 9299.24MiB. Peak reserved: 10858.00MiB +12/10/2024 01:04:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 2.55K | tokens_per_sec: 412K | tokens_per_sec_per_gpu: 6.43K | global_batch_size: 512 | lm_loss: 12.2 | lr: 0.00015 | model_tflops_per_gpu: 53.1 | hardware_tflops_per_gpu: 53.1 | grad_norm: 0.86 | cuda_memory_allocated: 4.03G | cuda_max_memory_reserved: 12.7G | hd_total_memory_tb: 312G | hd_used_memory_tb: 67.5G | hd_free_memory_tb: 245G +12/10/2024 01:04:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3842.57MiB. Peak allocated 6247.85MiB. Peak reserved: 12092.00MiB +12/10/2024 01:04:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3842.58MiB. Peak allocated 9447.33MiB. Peak reserved: 12098.00MiB +12/10/2024 01:04:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 569 | tokens_per_sec: 1.84M | tokens_per_sec_per_gpu: 28.8K | global_batch_size: 512 | lm_loss: 12.2 | lr: 0.0003 | model_tflops_per_gpu: 238 | hardware_tflops_per_gpu: 238 | grad_norm: 0.861 | cuda_memory_allocated: 4.03G | cuda_max_memory_reserved: 12.7G | hd_total_memory_tb: 312G | hd_used_memory_tb: 67.5G | hd_free_memory_tb: 245G +12/10/2024 01:04:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3842.57MiB. Peak allocated 6247.85MiB. Peak reserved: 12098.00MiB +12/10/2024 01:04:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 3842.58MiB. Peak allocated 9447.33MiB. Peak reserved: 12098.00MiB +num_params +{'total': 1241649152, 'local': 620824576}num_params +num_params +{'total': 1241649152, 'local': 620824576} + +num_params +{'total': 1241649152, 'local': 620824576} +num_paramsnum_paramsnum_params + +num_params{'total': 1241649152, 'local': 620824576}num_params{'total': 1241649152, 'local': 620824576} + + + + +num_params{'total': 1241649152, 'local': 620824576}num_params +num_params + +{'total': 1241649152, 'local': 620824576}{'total': 1241649152, 'local': 620824576} + +num_params +{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} +num_paramsnum_params + +{'total': 1241649152, 'local': 620824576} +num_params{'total': 1241649152, 'local': 620824576}num_params + +num_params +{'total': 1241649152, 'local': 620824576} +num_params +num_params{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576}{'total': 1241649152, 'local': 620824576} + +num_params +num_paramsnum_paramsnum_params + +{'total': 1241649152, 'local': 620824576}{'total': 1241649152, 'local': 620824576} + + +num_paramsnum_paramsnum_paramsnum_paramsnum_params{'total': 1241649152, 'local': 620824576} + + + +num_paramsnum_params + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576}num_params +num_params +num_params +{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576}num_params + +{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} +num_paramsnum_params +num_params + +{'total': 1241649152, 'local': 620824576}{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} + + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +num_params +num_paramsnum_paramsnum_params + +num_paramsnum_params +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +num_paramsnum_params +num_paramsnum_params + + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576}num_paramsnum_params +num_params + +{'total': 1241649152, 'local': 620824576} + + +{'total': 1241649152, 'local': 620824576}{'total': 1241649152, 'local': 620824576}{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +num_params + +{'total': 1241649152, 'local': 620824576} +num_params{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +num_paramsnum_params + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +num_paramsnum_params + +{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} + +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +num_params +{'total': 1241649152, 'local': 620824576} +{'total': 1241649152, 'local': 620824576} +12/10/2024 01:04:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 566 | tokens_per_sec: 1.85M | tokens_per_sec_per_gpu: 28.9K | global_batch_size: 512 | lm_loss: 12.2 | lr: 0.000296 | model_tflops_per_gpu: 239 | hardware_tflops_per_gpu: 239 | grad_norm: 0.849 | cuda_memory_allocated: 4.03G | cuda_max_memory_reserved: 12.7G | hd_total_memory_tb: 312G | hd_used_memory_tb: 67.5G | hd_free_memory_tb: 245G +num_params +{'total': 1241649152, 'local': 620824576} +12/10/2024 01:04:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/10/2024 01:04:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: | -------- | -------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | ----------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/10/2024 01:04:35 [INFO|DP=16|PP=0|TP=0|ip-26-0-166-244]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=25|PP=0|TP=0|ip-26-0-167-51]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=16|PP=0|TP=1|ip-26-0-166-244]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=5|PP=0|TP=0|ip-26-0-166-125]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=6|PP=0|TP=0|ip-26-0-166-125]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=4|PP=0|TP=0|ip-26-0-166-125]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=5|PP=0|TP=1|ip-26-0-166-125]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=9|PP=0|TP=0|ip-26-0-166-15]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=10|PP=0|TP=1|ip-26-0-166-15]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=27|PP=0|TP=1|ip-26-0-167-51]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=21|PP=0|TP=0|ip-26-0-166-36]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=20|PP=0|TP=1|ip-26-0-166-36]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=18|PP=0|TP=0|ip-26-0-166-244]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=19|PP=0|TP=1|ip-26-0-166-244]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=17|PP=0|TP=0|ip-26-0-166-244]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=18|PP=0|TP=1|ip-26-0-166-244]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=4|PP=0|TP=1|ip-26-0-166-125]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=13|PP=0|TP=0|ip-26-0-166-214]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=12|PP=0|TP=0|ip-26-0-166-214]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=24|PP=0|TP=1|ip-26-0-167-51]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=8|PP=0|TP=1|ip-26-0-166-15]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=10|PP=0|TP=0|ip-26-0-166-15]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=8|PP=0|TP=0|ip-26-0-166-15]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=11|PP=0|TP=1|ip-26-0-166-15]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=11|PP=0|TP=0|ip-26-0-166-15]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=26|PP=0|TP=0|ip-26-0-167-51]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=27|PP=0|TP=0|ip-26-0-167-51]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=26|PP=0|TP=1|ip-26-0-167-51]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=30|PP=0|TP=0|ip-26-0-167-9]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=20|PP=0|TP=0|ip-26-0-166-36]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=22|PP=0|TP=0|ip-26-0-166-36]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=23|PP=0|TP=0|ip-26-0-166-36]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=21|PP=0|TP=1|ip-26-0-166-36]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=19|PP=0|TP=0|ip-26-0-166-244]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=7|PP=0|TP=0|ip-26-0-166-125]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=14|PP=0|TP=0|ip-26-0-166-214]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=25|PP=0|TP=1|ip-26-0-167-51]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=22|PP=0|TP=1|ip-26-0-166-36]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=6|PP=0|TP=1|ip-26-0-166-125]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=7|PP=0|TP=1|ip-26-0-166-125]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=12|PP=0|TP=1|ip-26-0-166-214]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=13|PP=0|TP=1|ip-26-0-166-214]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=31|PP=0|TP=0|ip-26-0-167-9]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=23|PP=0|TP=1|ip-26-0-166-36]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=15|PP=0|TP=1|ip-26-0-166-214]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=15|PP=0|TP=0|ip-26-0-166-214]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=30|PP=0|TP=1|ip-26-0-167-9]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=28|PP=0|TP=1|ip-26-0-167-9]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=28|PP=0|TP=0|ip-26-0-167-9]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=29|PP=0|TP=1|ip-26-0-167-9]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=31|PP=0|TP=1|ip-26-0-167-9]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=29|PP=0|TP=0|ip-26-0-167-9]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=14|PP=0|TP=1|ip-26-0-166-214]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=9|PP=0|TP=1|ip-26-0-166-15]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=17|PP=0|TP=1|ip-26-0-166-244]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=24|PP=0|TP=0|ip-26-0-167-51]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: | 13265039 | 1.34G_dp32_tp2_pp1_acc8_mbs2_seq2048_zero1_tpmodeALL_vocab131k | 8 | 2048 | 2 | 8 | 512 | 238.92 | 238.92 | 28944.13 | 239.23 | 173.59 | 163.12 | 458.91 | 264.87 | 264.26 | 6.10 | 11.81 | 32 | 1 | 2 | | TensorParallelLinearMode.ALL_REDUCE | False | 2048 | silu | 16 | 32 | 8 | 2048 | 131072 | True | torch.bfloat16 | 1 | 25 | True | 1.24G | 621M | +12/10/2024 01:04:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=2|PP=0|TP=1|ip-26-0-160-225]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=1|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-225]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=3|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=3|PP=0|TP=1|ip-26-0-160-225]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=2|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/10/2024 01:04:35 [INFO|DP=1|PP=0|TP=1|ip-26-0-160-225]: Throughput logging complete +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-10 01:04:35,467] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1280100 closing signal SIGTERM +[2024-12-10 01:04:35,467] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1280101 closing signal SIGTERM +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1280102 closing signal SIGTERM +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22695 closing signal SIGTERM +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22696 closing signal SIGTERM +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22697 closing signal SIGTERM +[2024-12-10 01:04:35,468] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +slurmstepd: error: *** JOB 13265039 ON ip-26-0-160-225 CANCELLED AT 2024-12-10T01:04:35 *** +[2024-12-10 01:04:35,468] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1280103 closing signal SIGTERM +[2024-12-10 01:04:35,464] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22487 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1577047 closing signal SIGTERM +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22488 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1577048 closing signal SIGTERM +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22698 closing signal SIGTERM +[2024-12-10 01:04:35,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75238 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1577049 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1577050 closing signal SIGTERM +[2024-12-10 01:04:35,468] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 01:04:35,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75239 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1577051 closing signal SIGTERM +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22489 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 95115 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 95116 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 95117 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1577052 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 95118 closing signal SIGTERM +[2024-12-10 01:04:35,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22490 closing signal SIGTERM +[2024-12-10 01:04:35,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75240 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22699 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22700 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22491 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22492 closing signal SIGTERM +[2024-12-10 01:04:35,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22701 closing signal SIGTERM +[2024-12-10 01:04:35,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75241 closing signal SIGTERM +[2024-12-10 01:04:35,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22702 closing signal SIGTERM +[2024-12-10 01:04:35,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22493 closing signal SIGTERM +[2024-12-10 01:04:35,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75242 closing signal SIGTERM +[2024-12-10 01:04:35,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75243 closing signal SIGTERM +[2024-12-10 01:04:35,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 95119 closing signal SIGTERM +[2024-12-10 01:04:35,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 95120 closing signal SIGTERM +[2024-12-10 01:04:35,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 95121 closing signal SIGTERM +[2024-12-10 01:04:35,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75244 closing signal SIGTERM +[2024-12-10 01:04:35,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1280104 closing signal SIGTERM +[2024-12-10 01:04:35,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75245 closing signal SIGTERM +[2024-12-10 01:04:35,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1280105 closing signal SIGTERM +[2024-12-10 01:04:35,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1280106 closing signal SIGTERM +[2024-12-10 01:04:35,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1280107 closing signal SIGTERM +[2024-12-10 01:04:35,471] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1577053 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 932346 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 932347 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 932348 closing signal SIGTERM +[2024-12-10 01:04:35,471] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1577054 closing signal SIGTERM +[2024-12-10 01:04:35,469] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 932349 closing signal SIGTERM +[2024-12-10 01:04:35,471] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22494 closing signal SIGTERM +[2024-12-10 01:04:35,472] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 932350 closing signal SIGTERM +[2024-12-10 01:04:35,472] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 932351 closing signal SIGTERM +[2024-12-10 01:04:35,472] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 95122 closing signal SIGTERM +[2024-12-10 01:04:35,472] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 932352 closing signal SIGTERM diff --git a/logs/13265359-bench_1.34G_dp2_tp4_pp1_acc2_mbs128_seq2048_zero1_tpmodeALL_vocab131k.out b/logs/13265359-bench_1.34G_dp2_tp4_pp1_acc2_mbs128_seq2048_zero1_tpmodeALL_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..2cabf58ffca725becfa3f161dbf5a38214f8b35e --- /dev/null +++ b/logs/13265359-bench_1.34G_dp2_tp4_pp1_acc2_mbs128_seq2048_zero1_tpmodeALL_vocab131k.out @@ -0,0 +1,760 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-171-230 ++ export NODELIST=ip-26-0-171-230 ++ NODELIST=ip-26-0-171-230 +++ scontrol show hostnames ip-26-0-171-230 +++ head -n1 ++ export MASTER_NODE=ip-26-0-171-230 ++ MASTER_NODE=ip-26-0-171-230 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-171-230' +Master node: ip-26-0-171-230 ++ echo 'All nodes: ip-26-0-171-230' +All nodes: ip-26-0-171-230 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13265359 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-171-230:12356 run_train.py --config-file benchmark/configs/config_1.34G_dp2_tp4_pp1_acc2_mbs128_seq2048_zero1_tpmodeALL_vocab131k.yaml +[2024-12-10 01:01:06,572] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 01:01:06,572] torch.distributed.run: [WARNING] +[2024-12-10 01:01:06,572] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 01:01:06,572] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 01:01:06,572] torch.distributed.run: [WARNING] ***************************************** +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: Config: +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: Config(general=GeneralArgs(project='debug', +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: run='1.34G_dp2_tp4_pp1_acc2_mbs128_seq2048_zero1_tpmodeALL_vocab131k', +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: seed=42, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: step=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: consumed_train_samples=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: ignore_sanity_checks=True), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: parallelism=ParallelismArgs(dp=2, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: pp=1, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: tp=4, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: pp_engine=, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: tp_mode=, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: tp_linear_async_communication=False, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: recompute_layer=False, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: tp_recompute_allgather=True, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: expert_parallel_size=1), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: eos_token_id=0, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: hidden_act='silu', +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: hidden_size=2048, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: initializer_range=0.02, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: intermediate_size=8192, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: is_llama_config=True, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: max_position_embeddings=2048, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: num_attention_heads=32, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: num_hidden_layers=16, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: num_key_value_heads=8, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: pad_token_id=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: pretraining_tp=1, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: rms_norm_eps=1e-05, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: rope_scaling=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: rope_theta=10000.0, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: rope_interleaved=False, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: tie_word_embeddings=True, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: use_cache=True, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: vocab_size=131072), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: init_method=RandomInit(std=0.02), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: dtype=torch.bfloat16, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: make_vocab_size_divisible_by=1, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: ddp_bucket_cap_mb=25), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: tokenizer_revision=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: tokenizer_max_length=None), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: checkpoint_interval=10000, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: save_initial_state=False, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: save_final_state=False, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: resume_checkpoint_path=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: checkpoints_path_is_shared_file_system=False), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: logging=LoggingArgs(log_level='info', +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: log_level_replica='info', +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: iteration_step_info_interval=1), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: tokens=TokensArgs(sequence_length=2048, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: train_steps=100, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: micro_batch_size=128, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: batch_accumulation_per_replica=2, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: val_check_interval=100, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: limit_val_batches=0, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: limit_test_batches=0), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: adam_beta1=0.9, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: adam_beta2=0.95, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: torch_adam_is_fused=True, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: name='adamW'), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: zero_stage=1, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: weight_decay=0.01, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: clip_grad=1.0, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: accumulate_grad_in_fp32=True, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: lr_warmup_steps=2, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: lr_warmup_style='linear', +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: lr_decay_style='cosine', +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: lr_decay_steps=13, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: lr_decay_starting_step=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: min_decay_lr=1e-05)), +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: start_training_step=1, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: data=DataArgs(dataset=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: seed=42, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: num_loading_workers=1))], +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: profiler=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: lighteval=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: s3_upload=None) +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: Model Config: +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: LlamaConfig(bos_token_id=0, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: eos_token_id=0, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: hidden_act='silu', +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: hidden_size=2048, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: initializer_range=0.02, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: intermediate_size=8192, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: is_llama_config=True, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: max_position_embeddings=2048, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: num_attention_heads=32, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: num_hidden_layers=16, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: num_key_value_heads=8, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: pad_token_id=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: pretraining_tp=1, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: rms_norm_eps=1e-05, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: rope_scaling=None, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: rope_theta=10000.0, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: rope_interleaved=False, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: tie_word_embeddings=True, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: use_cache=True, +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: vocab_size=131072) +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: Building model.. +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: Initialize RoPE Theta = 10000.0 +12/10/2024 01:01:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: Setting PP block ranks... +12/10/2024 01:01:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: Total number of parameters: 1.24G (2368.52MiB) +12/10/2024 01:01:46 [INFO|DP=0|PP=0|TP=2|ip-26-0-171-230]: Local number of parameters: 310M (592.13MiB) +12/10/2024 01:01:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: Local number of parameters: 310M (592.13MiB) +12/10/2024 01:01:46 [INFO|DP=0|PP=0|TP=1|ip-26-0-171-230]: Local number of parameters: 310M (592.13MiB) +12/10/2024 01:01:46 [INFO|DP=0|PP=0|TP=3|ip-26-0-171-230]: Local number of parameters: 310M (592.13MiB) +12/10/2024 01:01:46 [INFO|DP=0|PP=0|TP=2|ip-26-0-171-230]: [After model building] Memory usage: 592.15MiB. Peak allocated: 5632.00MiB Peak reserved: 15874.00MiB +12/10/2024 01:01:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: [After model building] Memory usage: 592.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/10/2024 01:01:46 [INFO|DP=0|PP=0|TP=1|ip-26-0-171-230]: [After model building] Memory usage: 592.15MiB. Peak allocated: 5632.00MiB Peak reserved: 17922.00MiB +12/10/2024 01:01:46 [INFO|DP=0|PP=0|TP=3|ip-26-0-171-230]: [After model building] Memory usage: 592.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/10/2024 01:01:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: No checkpoint path provided. +12/10/2024 01:01:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: Parametrizing model parameters using StandardParametrizator +12/10/2024 01:01:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: [Optimizer Building] Using LearningRateForSP as learning rate +12/10/2024 01:01:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: [ZeRO sharding] Size of optimizer params per rank: +12/10/2024 01:01:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: [ZeRO sharding] DP Rank 0 has 155M out of 310M (50.00%) params' optimizer states +12/10/2024 01:01:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: [ZeRO sharding] DP Rank 1 has 155M out of 310M (50.00%) params' optimizer states +12/10/2024 01:01:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/10/2024 01:01:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: Using dummy data generator +12/10/2024 01:01:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: [Training Plan] There are 1 training stages +12/10/2024 01:01:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: [Stage Stable Training Stage] start from step 1 +12/10/2024 01:01:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: +12/10/2024 01:01:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-230]: [Start training] datetime: 2024-12-10 01:01:49.273077 | mbs: 128 | grad_accum: 2 | global_batch_size: 512 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) +sharded_logits = self.model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)sharded_logits = self.model( + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 690, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.o_proj(attention_output) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 595, in row_linear + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + out = F.linear(input, weight, bias) + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1024.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 801.94 MiB is free. Including non-PyTorch memory, this process has 78.54 GiB memory in use. Of the allocated memory 67.80 GiB is allocated by PyTorch, and 317.69 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 690, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.o_proj(attention_output) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 595, in row_linear + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + out = F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1024.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 929.94 MiB is free. Including non-PyTorch memory, this process has 78.41 GiB memory in use. Of the allocated memory 67.80 GiB is allocated by PyTorch, and 189.69 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 690, in forward + output = self.o_proj(attention_output) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 595, in row_linear + out = F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1024.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 977.94 MiB is free. Including non-PyTorch memory, this process has 78.36 GiB memory in use. Of the allocated memory 67.80 GiB is allocated by PyTorch, and 189.69 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[2024-12-10 01:01:51,666] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 24991 closing signal SIGTERM +[2024-12-10 01:01:51,667] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 24992 closing signal SIGTERM +[2024-12-10 01:01:51,667] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 24993 closing signal SIGTERM +[2024-12-10 01:01:51,667] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 24994 closing signal SIGTERM +[2024-12-10 01:01:51,668] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 24996 closing signal SIGTERM +[2024-12-10 01:01:51,668] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 24997 closing signal SIGTERM +[2024-12-10 01:01:51,669] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 24998 closing signal SIGTERM +[2024-12-10 01:01:53,149] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 4 (pid: 24995) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-10_01:01:51 + host : ip-26-0-171-230.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 24995) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-171-230: task 0: Exited with exit code 1 diff --git a/logs/13265594-bench_1.34G_dp4_tp4_pp1_acc4_mbs32_seq8192_zero1_tpmodeRED_vocab131k.out b/logs/13265594-bench_1.34G_dp4_tp4_pp1_acc4_mbs32_seq8192_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..78571fbffd2463e62da1322e430bbeb12bf4f21d --- /dev/null +++ b/logs/13265594-bench_1.34G_dp4_tp4_pp1_acc4_mbs32_seq8192_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,1225 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-165-38,ip-26-0-166-15 ++ export 'NODELIST=ip-26-0-165-38 +ip-26-0-166-15' ++ NODELIST='ip-26-0-165-38 +ip-26-0-166-15' +++ scontrol show hostnames ip-26-0-165-38,ip-26-0-166-15 +++ head -n1 ++ export MASTER_NODE=ip-26-0-165-38 ++ MASTER_NODE=ip-26-0-165-38 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-165-38' +Master node: ip-26-0-165-38 ++ echo 'All nodes: ip-26-0-165-38 +ip-26-0-166-15' +All nodes: ip-26-0-165-38 +ip-26-0-166-15 ++ echo 'World size: 16' +World size: 16 ++ srun torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13265594 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-165-38:12356 run_train.py --config-file benchmark/configs/config_1.34G_dp4_tp4_pp1_acc4_mbs32_seq8192_zero1_tpmodeRED_vocab131k.yaml +[2024-12-10 02:56:06,791] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 02:56:06,791] torch.distributed.run: [WARNING] +[2024-12-10 02:56:06,791] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 02:56:06,791] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 02:56:06,791] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 02:56:07,059] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 02:56:07,060] torch.distributed.run: [WARNING] +[2024-12-10 02:56:07,060] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 02:56:07,060] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 02:56:07,060] torch.distributed.run: [WARNING] ***************************************** +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: Config: +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: Config(general=GeneralArgs(project='debug', +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: run='1.34G_dp4_tp4_pp1_acc4_mbs32_seq8192_zero1_tpmodeRED_vocab131k', +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: seed=42, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: step=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: consumed_train_samples=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: ignore_sanity_checks=True), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: parallelism=ParallelismArgs(dp=4, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: pp=1, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: tp=4, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: pp_engine=, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: tp_mode=, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: tp_linear_async_communication=True, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: recompute_layer=False, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: tp_recompute_allgather=True, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: expert_parallel_size=1), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: eos_token_id=0, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: hidden_act='silu', +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: hidden_size=2048, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: initializer_range=0.02, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: intermediate_size=8192, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: is_llama_config=True, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: max_position_embeddings=8192, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: num_attention_heads=32, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: num_hidden_layers=16, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: num_key_value_heads=8, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: pad_token_id=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: pretraining_tp=1, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: rms_norm_eps=1e-05, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: rope_scaling=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: rope_theta=10000.0, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: rope_interleaved=False, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: tie_word_embeddings=True, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: use_cache=True, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: vocab_size=131072), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: init_method=RandomInit(std=0.02), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: dtype=torch.bfloat16, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: make_vocab_size_divisible_by=1, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: ddp_bucket_cap_mb=25), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: tokenizer_revision=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: tokenizer_max_length=None), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: checkpoint_interval=10000, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: save_initial_state=False, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: save_final_state=False, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: resume_checkpoint_path=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: checkpoints_path_is_shared_file_system=False), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: logging=LoggingArgs(log_level='info', +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: log_level_replica='info', +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: iteration_step_info_interval=1), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: tokens=TokensArgs(sequence_length=8192, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: train_steps=100, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: micro_batch_size=32, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: batch_accumulation_per_replica=4, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: val_check_interval=100, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: limit_val_batches=0, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: limit_test_batches=0), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: adam_beta1=0.9, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: adam_beta2=0.95, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: torch_adam_is_fused=True, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: name='adamW'), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: zero_stage=1, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: weight_decay=0.01, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: clip_grad=1.0, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: accumulate_grad_in_fp32=True, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: lr_warmup_steps=2, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: lr_warmup_style='linear', +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: lr_decay_style='cosine', +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: lr_decay_steps=13, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: lr_decay_starting_step=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: min_decay_lr=1e-05)), +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: start_training_step=1, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: data=DataArgs(dataset=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: seed=42, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: num_loading_workers=1))], +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: profiler=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: lighteval=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: s3_upload=None) +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: Model Config: +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: LlamaConfig(bos_token_id=0, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: eos_token_id=0, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: hidden_act='silu', +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: hidden_size=2048, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: initializer_range=0.02, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: intermediate_size=8192, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: is_llama_config=True, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: max_position_embeddings=8192, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: num_attention_heads=32, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: num_hidden_layers=16, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: num_key_value_heads=8, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: pad_token_id=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: pretraining_tp=1, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: rms_norm_eps=1e-05, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: rope_scaling=None, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: rope_theta=10000.0, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: rope_interleaved=False, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: tie_word_embeddings=True, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: use_cache=True, +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: vocab_size=131072) +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: Building model.. +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: Initialize RoPE Theta = 10000.0 +12/10/2024 02:56:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: Setting PP block ranks... +12/10/2024 02:56:53 [INFO|DP=0|PP=0|TP=3|ip-26-0-165-38]: Local number of parameters: 310M (592.13MiB) +12/10/2024 02:56:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: Total number of parameters: 1.24G (2368.52MiB) +12/10/2024 02:56:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: Local number of parameters: 310M (592.13MiB) +12/10/2024 02:56:53 [INFO|DP=0|PP=0|TP=1|ip-26-0-165-38]: Local number of parameters: 310M (592.13MiB) +12/10/2024 02:56:53 [INFO|DP=0|PP=0|TP=2|ip-26-0-165-38]: Local number of parameters: 310M (592.13MiB) +12/10/2024 02:56:53 [INFO|DP=0|PP=0|TP=3|ip-26-0-165-38]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 02:56:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 02:56:53 [INFO|DP=0|PP=0|TP=1|ip-26-0-165-38]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 02:56:53 [INFO|DP=0|PP=0|TP=2|ip-26-0-165-38]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 02:56:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: No checkpoint path provided. +12/10/2024 02:56:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: Parametrizing model parameters using StandardParametrizator +12/10/2024 02:56:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: [Optimizer Building] Using LearningRateForSP as learning rate +12/10/2024 02:56:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: [ZeRO sharding] Size of optimizer params per rank: +12/10/2024 02:56:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: [ZeRO sharding] DP Rank 0 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 02:56:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: [ZeRO sharding] DP Rank 1 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 02:56:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: [ZeRO sharding] DP Rank 2 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 02:56:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: [ZeRO sharding] DP Rank 3 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 02:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/10/2024 02:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: Using dummy data generator +12/10/2024 02:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: [Training Plan] There are 1 training stages +12/10/2024 02:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: [Stage Stable Training Stage] start from step 1 +12/10/2024 02:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: +12/10/2024 02:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-38]: [Start training] datetime: 2024-12-10 02:56:56.170264 | mbs: 32 | grad_accum: 4 | global_batch_size: 512 | sequence_length: 8192 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 690, in forward + output = self.o_proj(attention_output) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 466, in forward + out = F.linear(tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1024.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 831.94 MiB is free. Including non-PyTorch memory, this process has 78.51 GiB memory in use. Of the allocated memory 66.80 GiB is allocated by PyTorch, and 1.25 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward +merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + return self._call_impl(*args, **kwargs) +return func(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 1.49 GiB is free. Including non-PyTorch memory, this process has 77.83 GiB memory in use. Of the allocated memory 66.93 GiB is allocated by PyTorch, and 899.76 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs)gathered_output = torch.empty( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 1.44 GiB is free. Including non-PyTorch memory, this process has 77.88 GiB memory in use. Of the allocated memory 66.93 GiB is allocated by PyTorch, and 515.76 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 1.19 GiB is free. Including non-PyTorch memory, this process has 78.13 GiB memory in use. Of the allocated memory 66.93 GiB is allocated by PyTorch, and 771.76 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 1.59 GiB is free. Including non-PyTorch memory, this process has 77.72 GiB memory in use. Of the allocated memory 66.93 GiB is allocated by PyTorch, and 387.76 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + output = model(**micro_batch) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 1.06 GiB is free. Including non-PyTorch memory, this process has 78.26 GiB memory in use. Of the allocated memory 66.93 GiB is allocated by PyTorch, and 899.76 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + gathered_output = torch.empty( + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 1.09 GiB is free. Including non-PyTorch memory, this process has 78.22 GiB memory in use. Of the allocated memory 66.93 GiB is allocated by PyTorch, and 899.76 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 1.37 GiB is free. Including non-PyTorch memory, this process has 77.95 GiB memory in use. Of the allocated memory 66.93 GiB is allocated by PyTorch, and 1.00 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 1.06 GiB is free. Including non-PyTorch memory, this process has 78.26 GiB memory in use. Of the allocated memory 66.93 GiB is allocated by PyTorch, and 899.76 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[2024-12-10 02:57:03,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 730984 closing signal SIGTERM +[2024-12-10 02:57:03,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 730985 closing signal SIGTERM +[2024-12-10 02:57:03,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 730986 closing signal SIGTERM +[2024-12-10 02:57:03,223] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 730987 closing signal SIGTERM +[2024-12-10 02:57:03,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 63188 closing signal SIGTERM +[2024-12-10 02:57:03,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 63190 closing signal SIGTERM +[2024-12-10 02:57:03,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 63191 closing signal SIGTERM +[2024-12-10 02:57:06,579] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 1 (pid: 63189) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-10_02:57:03 + host : ip-26-0-166-15.ec2.internal + rank : 12 (local_rank: 4) + exitcode : 1 (pid: 63192) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-10_02:57:03 + host : ip-26-0-166-15.ec2.internal + rank : 13 (local_rank: 5) + exitcode : 1 (pid: 63193) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-10_02:57:03 + host : ip-26-0-166-15.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 63194) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-10_02:57:03 + host : ip-26-0-166-15.ec2.internal + rank : 15 (local_rank: 7) + exitcode : 1 (pid: 63195) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-10_02:57:03 + host : ip-26-0-166-15.ec2.internal + rank : 9 (local_rank: 1) + exitcode : 1 (pid: 63189) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-10 02:57:06,719] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 4 (pid: 730988) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-10_02:57:03 + host : ip-26-0-165-38.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 730989) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-10_02:57:03 + host : ip-26-0-165-38.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 730990) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-10_02:57:03 + host : ip-26-0-165-38.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 730991) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-10_02:57:03 + host : ip-26-0-165-38.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 730988) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-166-15: task 1: Exited with exit code 1 +srun: error: ip-26-0-165-38: task 0: Exited with exit code 1 diff --git a/logs/13265599-bench_1.34G_dp4_tp4_pp1_acc2_mbs64_seq8192_zero1_tpmodeALL_vocab131k.out b/logs/13265599-bench_1.34G_dp4_tp4_pp1_acc2_mbs64_seq8192_zero1_tpmodeALL_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..bd0a31d4b66ca64af85790fcfaac5d50fa1acea2 --- /dev/null +++ b/logs/13265599-bench_1.34G_dp4_tp4_pp1_acc2_mbs64_seq8192_zero1_tpmodeALL_vocab131k.out @@ -0,0 +1,1577 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-164-[187,236]' ++ export 'NODELIST=ip-26-0-164-187 +ip-26-0-164-236' ++ NODELIST='ip-26-0-164-187 +ip-26-0-164-236' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-164-[187,236]' ++ export MASTER_NODE=ip-26-0-164-187 ++ MASTER_NODE=ip-26-0-164-187 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-164-187' +Master node: ip-26-0-164-187 ++ echo 'All nodes: ip-26-0-164-187 +ip-26-0-164-236' +All nodes: ip-26-0-164-187 +ip-26-0-164-236 ++ echo 'World size: 16' +World size: 16 ++ srun torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13265599 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-164-187:12356 run_train.py --config-file benchmark/configs/config_1.34G_dp4_tp4_pp1_acc2_mbs64_seq8192_zero1_tpmodeALL_vocab131k.yaml +[2024-12-10 02:59:40,965] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 02:59:40,965] torch.distributed.run: [WARNING] +[2024-12-10 02:59:40,965] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 02:59:40,965] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 02:59:40,965] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 02:59:41,166] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 02:59:41,166] torch.distributed.run: [WARNING] +[2024-12-10 02:59:41,166] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 02:59:41,166] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 02:59:41,166] torch.distributed.run: [WARNING] ***************************************** +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Config: +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Config(general=GeneralArgs(project='debug', +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: run='1.34G_dp4_tp4_pp1_acc2_mbs64_seq8192_zero1_tpmodeALL_vocab131k', +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: seed=42, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: step=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: consumed_train_samples=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: ignore_sanity_checks=True), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: parallelism=ParallelismArgs(dp=4, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: pp=1, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: tp=4, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: pp_engine=, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: tp_mode=, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: tp_linear_async_communication=False, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: recompute_layer=False, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: tp_recompute_allgather=True, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: expert_parallel_size=1), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: eos_token_id=0, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: hidden_act='silu', +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: hidden_size=2048, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: initializer_range=0.02, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: intermediate_size=8192, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: is_llama_config=True, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: max_position_embeddings=8192, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: num_attention_heads=32, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: num_hidden_layers=16, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: num_key_value_heads=8, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: pad_token_id=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: pretraining_tp=1, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: rms_norm_eps=1e-05, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: rope_scaling=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: rope_theta=10000.0, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: rope_interleaved=False, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: tie_word_embeddings=True, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: use_cache=True, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: vocab_size=131072), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: init_method=RandomInit(std=0.02), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: dtype=torch.bfloat16, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: make_vocab_size_divisible_by=1, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: ddp_bucket_cap_mb=25), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: tokenizer_revision=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: tokenizer_max_length=None), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: checkpoint_interval=10000, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: save_initial_state=False, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: save_final_state=False, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: resume_checkpoint_path=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: checkpoints_path_is_shared_file_system=False), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: logging=LoggingArgs(log_level='info', +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: log_level_replica='info', +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: iteration_step_info_interval=1), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: tokens=TokensArgs(sequence_length=8192, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: train_steps=100, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: micro_batch_size=64, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: batch_accumulation_per_replica=2, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: val_check_interval=100, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: limit_val_batches=0, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: limit_test_batches=0), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: adam_beta1=0.9, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: adam_beta2=0.95, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: torch_adam_is_fused=True, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: name='adamW'), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: zero_stage=1, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: weight_decay=0.01, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: clip_grad=1.0, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: accumulate_grad_in_fp32=True, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: lr_warmup_steps=2, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: lr_warmup_style='linear', +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: lr_decay_style='cosine', +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: lr_decay_steps=13, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: lr_decay_starting_step=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: min_decay_lr=1e-05)), +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: start_training_step=1, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: data=DataArgs(dataset=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: seed=42, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: num_loading_workers=1))], +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: profiler=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: lighteval=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: s3_upload=None) +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Model Config: +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: LlamaConfig(bos_token_id=0, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: eos_token_id=0, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: hidden_act='silu', +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: hidden_size=2048, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: initializer_range=0.02, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: intermediate_size=8192, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: is_llama_config=True, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: max_position_embeddings=8192, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: num_attention_heads=32, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: num_hidden_layers=16, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: num_key_value_heads=8, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: pad_token_id=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: pretraining_tp=1, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: rms_norm_eps=1e-05, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: rope_scaling=None, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: rope_theta=10000.0, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: rope_interleaved=False, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: tie_word_embeddings=True, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: use_cache=True, +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: vocab_size=131072) +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Building model.. +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Initialize RoPE Theta = 10000.0 +12/10/2024 03:00:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Setting PP block ranks... +12/10/2024 03:00:24 [INFO|DP=0|PP=0|TP=3|ip-26-0-164-187]: Local number of parameters: 310M (592.13MiB) +12/10/2024 03:00:24 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Total number of parameters: 1.24G (2368.52MiB) +12/10/2024 03:00:24 [INFO|DP=0|PP=0|TP=2|ip-26-0-164-187]: Local number of parameters: 310M (592.13MiB) +12/10/2024 03:00:24 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Local number of parameters: 310M (592.13MiB) +12/10/2024 03:00:24 [INFO|DP=0|PP=0|TP=1|ip-26-0-164-187]: Local number of parameters: 310M (592.13MiB) +12/10/2024 03:00:24 [INFO|DP=0|PP=0|TP=3|ip-26-0-164-187]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 03:00:24 [INFO|DP=0|PP=0|TP=2|ip-26-0-164-187]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 03:00:24 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 03:00:24 [INFO|DP=0|PP=0|TP=1|ip-26-0-164-187]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 03:00:24 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: No checkpoint path provided. +12/10/2024 03:00:24 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Parametrizing model parameters using StandardParametrizator +12/10/2024 03:00:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: [Optimizer Building] Using LearningRateForSP as learning rate +12/10/2024 03:00:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: [ZeRO sharding] Size of optimizer params per rank: +12/10/2024 03:00:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: [ZeRO sharding] DP Rank 0 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 03:00:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: [ZeRO sharding] DP Rank 1 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 03:00:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: [ZeRO sharding] DP Rank 2 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 03:00:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: [ZeRO sharding] DP Rank 3 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 03:00:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/10/2024 03:00:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Using dummy data generator +12/10/2024 03:00:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: [Training Plan] There are 1 training stages +12/10/2024 03:00:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: [Stage Stable Training Stage] start from step 1 +12/10/2024 03:00:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: +12/10/2024 03:00:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: [Start training] datetime: 2024-12-10 03:00:27.358105 | mbs: 64 | grad_accum: 2 | global_batch_size: 512 | sequence_length: 8192 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + outputs = self.pipeline_engine.train_batch_iter(return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 730, in _core_forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 730, in _core_forward + hidden_states = self.post_attention_layernorm(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.post_attention_layernorm(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/layer_norm.py", line 44, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/layer_norm.py", line 44, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return layer_norm_fn( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 875, in layer_norm_fn + return layer_norm_fn( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 875, in layer_norm_fn + return LayerNormFn.apply( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return LayerNormFn.apply( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 748, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 748, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + y, y1, mean, rstd, residual_out, seeds, dropout_mask, dropout_mask1 = _layer_norm_fwd( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 297, in _layer_norm_fwd + y, y1, mean, rstd, residual_out, seeds, dropout_mask, dropout_mask1 = _layer_norm_fwd( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 297, in _layer_norm_fwd + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + y = torch.empty_like(x, dtype=x.dtype if out_dtype is None else out_dtype) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 1.86 GiB is free. Including non-PyTorch memory, this process has 77.46 GiB memory in use. Of the allocated memory 65.15 GiB is allocated by PyTorch, and 1.89 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + y = torch.empty_like(x, dtype=x.dtype if out_dtype is None else out_dtype) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 1.08 GiB is free. Including non-PyTorch memory, this process has 78.24 GiB memory in use. Of the allocated memory 65.15 GiB is allocated by PyTorch, and 2.64 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 730, in _core_forward + hidden_states = self.post_attention_layernorm(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/layer_norm.py", line 44, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return layer_norm_fn( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 875, in layer_norm_fn + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return LayerNormFn.apply( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 748, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + y, y1, mean, rstd, residual_out, seeds, dropout_mask, dropout_mask1 = _layer_norm_fwd( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 297, in _layer_norm_fwd + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + y = torch.empty_like(x, dtype=x.dtype if out_dtype is None else out_dtype) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 1.83 GiB is free. Including non-PyTorch memory, this process has 77.49 GiB memory in use. Of the allocated memory 65.15 GiB is allocated by PyTorch, and 1.89 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 5.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 2.14 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) +output = model(**micro_batch) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + merged_states = self.gate_up_proj(hidden_states) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 845.94 MiB is free. Including non-PyTorch memory, this process has 78.49 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 909.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFreturn self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return column_linear(return F.linear(input, weight, bias) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 877.94 MiB is free. Including non-PyTorch memory, this process has 78.46 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 909.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 1.26 GiB is free. Including non-PyTorch memory, this process has 78.06 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 909.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 845.94 MiB is free. Including non-PyTorch memory, this process has 78.49 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 909.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 730, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + hidden_states = self.post_attention_layernorm(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/layer_norm.py", line 44, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return layer_norm_fn( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 875, in layer_norm_fn + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return LayerNormFn.apply( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 748, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + y, y1, mean, rstd, residual_out, seeds, dropout_mask, dropout_mask1 = _layer_norm_fwd( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/flash_attn/ops/triton/layer_norm.py", line 297, in _layer_norm_fwd +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + y = torch.empty_like(x, dtype=x.dtype if out_dtype is None else out_dtype) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 2.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 1.94 GiB is free. Including non-PyTorch memory, this process has 77.38 GiB memory in use. Of the allocated memory 65.15 GiB is allocated by PyTorch, and 1.89 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + sharded_logits = self.model(return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask)return F.linear(input, weight, bias) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 845.94 MiB is free. Including non-PyTorch memory, this process has 78.49 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 909.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 845.94 MiB is free. Including non-PyTorch memory, this process has 78.49 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 909.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 925.94 MiB is free. Including non-PyTorch memory, this process has 78.41 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 909.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/10/2024 03:00:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/10/2024 03:00:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-187]: Memory usage: 2075.48MiB. Peak allocated 5504.00MiB. Peak reserved: 36388.00MiB +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 250, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 669.94 MiB is free. Including non-PyTorch memory, this process has 78.66 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 1.14 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer.train(dataloader)output = model(**micro_batch) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 456, in train + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 493, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 278, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 44, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 1.08 GiB is free. Including non-PyTorch memory, this process has 78.24 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 653.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 1.19 GiB is free. Including non-PyTorch memory, this process has 78.12 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 653.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 445, in column_linear + return F.linear(input, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 845.94 MiB is free. Including non-PyTorch memory, this process has 78.49 GiB memory in use. Of the allocated memory 67.15 GiB is allocated by PyTorch, and 909.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[2024-12-10 03:00:37,390] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2770641) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-10 03:00:37,394] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3842799 closing signal SIGTERM +[2024-12-10 03:00:37,394] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3842800 closing signal SIGTERM +[2024-12-10 03:00:37,394] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3842801 closing signal SIGTERM +[2024-12-10 03:00:37,394] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3842802 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-236.ec2.internal + rank : 9 (local_rank: 1) + exitcode : 1 (pid: 2770642) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-236.ec2.internal + rank : 10 (local_rank: 2) + exitcode : 1 (pid: 2770643) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-236.ec2.internal + rank : 11 (local_rank: 3) + exitcode : 1 (pid: 2770644) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-236.ec2.internal + rank : 12 (local_rank: 4) + exitcode : 1 (pid: 2770645) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-236.ec2.internal + rank : 13 (local_rank: 5) + exitcode : 1 (pid: 2770646) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-236.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 2770647) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-236.ec2.internal + rank : 15 (local_rank: 7) + exitcode : 1 (pid: 2770648) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-236.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 2770641) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-164-236: task 1: Exited with exit code 1 +wandb: You can sync this run to the cloud by running: +wandb: wandb sync /tmp/wandb/offline-run-20241210_030029-lqc2h7aa +wandb: Find logs at: /tmp/wandb/offline-run-20241210_030029-lqc2h7aa/logs +[2024-12-10 03:00:40,992] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 4 (pid: 3842803) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-187.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 3842804) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-187.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 3842805) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-187.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 3842806) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-10_03:00:37 + host : ip-26-0-164-187.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 3842803) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-164-187: task 0: Exited with exit code 1 diff --git a/logs/13265609-bench_1.14G_dp4_tp4_pp1_acc32_mbs1_seq32768_zero1_tpmodeALL_vocab32k.out b/logs/13265609-bench_1.14G_dp4_tp4_pp1_acc32_mbs1_seq32768_zero1_tpmodeALL_vocab32k.out new file mode 100644 index 0000000000000000000000000000000000000000..9427edcbdc2b0bd4ca873d8376d57a7759f41f55 --- /dev/null +++ b/logs/13265609-bench_1.14G_dp4_tp4_pp1_acc32_mbs1_seq32768_zero1_tpmodeALL_vocab32k.out @@ -0,0 +1,637 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-164-236,ip-26-0-165-38 ++ export 'NODELIST=ip-26-0-164-236 +ip-26-0-165-38' ++ NODELIST='ip-26-0-164-236 +ip-26-0-165-38' +++ scontrol show hostnames ip-26-0-164-236,ip-26-0-165-38 +++ head -n1 ++ export MASTER_NODE=ip-26-0-164-236 ++ MASTER_NODE=ip-26-0-164-236 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-164-236' +Master node: ip-26-0-164-236 ++ echo 'All nodes: ip-26-0-164-236 +ip-26-0-165-38' +All nodes: ip-26-0-164-236 +ip-26-0-165-38 ++ echo 'World size: 16' +World size: 16 ++ srun torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13265609 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-164-236:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp4_pp1_acc32_mbs1_seq32768_zero1_tpmodeALL_vocab32k.yaml +[2024-12-10 03:04:18,827] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 03:04:18,865] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 03:04:18,827] torch.distributed.run: [WARNING] +[2024-12-10 03:04:18,827] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 03:04:18,827] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 03:04:18,827] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 03:04:18,865] torch.distributed.run: [WARNING] +[2024-12-10 03:04:18,865] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 03:04:18,865] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 03:04:18,865] torch.distributed.run: [WARNING] ***************************************** +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Config: +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Config(general=GeneralArgs(project='debug', +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: run='1.14G_dp4_tp4_pp1_acc32_mbs1_seq32768_zero1_tpmodeALL_vocab32k', +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: seed=42, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: step=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: consumed_train_samples=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: ignore_sanity_checks=True), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: parallelism=ParallelismArgs(dp=4, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: pp=1, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: tp=4, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: pp_engine=, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: tp_mode=, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: tp_linear_async_communication=False, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: recompute_layer=False, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: tp_recompute_allgather=True, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: expert_parallel_size=1), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: eos_token_id=0, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: hidden_act='silu', +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: hidden_size=2048, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: initializer_range=0.02, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: intermediate_size=8192, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: is_llama_config=True, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: max_position_embeddings=32768, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: num_attention_heads=32, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: num_hidden_layers=16, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: num_key_value_heads=8, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: pad_token_id=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: pretraining_tp=1, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: rms_norm_eps=1e-05, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: rope_scaling=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: rope_theta=10000.0, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: rope_interleaved=False, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: tie_word_embeddings=True, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: use_cache=True, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: vocab_size=32768), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: init_method=RandomInit(std=0.02), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: dtype=torch.bfloat16, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: make_vocab_size_divisible_by=1, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: ddp_bucket_cap_mb=25), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: tokenizer_revision=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: tokenizer_max_length=None), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: checkpoint_interval=10000, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: save_initial_state=False, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: save_final_state=False, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: resume_checkpoint_path=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: checkpoints_path_is_shared_file_system=False), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: logging=LoggingArgs(log_level='info', +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: log_level_replica='info', +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: iteration_step_info_interval=1), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: tokens=TokensArgs(sequence_length=32768, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: train_steps=100, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: micro_batch_size=1, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: batch_accumulation_per_replica=32, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: val_check_interval=100, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: limit_val_batches=0, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: limit_test_batches=0), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: adam_beta1=0.9, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: adam_beta2=0.95, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: torch_adam_is_fused=True, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: name='adamW'), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: zero_stage=1, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: weight_decay=0.01, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: clip_grad=1.0, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: accumulate_grad_in_fp32=True, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: lr_warmup_steps=2, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: lr_warmup_style='linear', +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: lr_decay_style='cosine', +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: lr_decay_steps=13, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: lr_decay_starting_step=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: min_decay_lr=1e-05)), +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: start_training_step=1, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: data=DataArgs(dataset=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: seed=42, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: num_loading_workers=1))], +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: profiler=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: lighteval=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: s3_upload=None) +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Model Config: +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: LlamaConfig(bos_token_id=0, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: eos_token_id=0, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: hidden_act='silu', +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: hidden_size=2048, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: initializer_range=0.02, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: intermediate_size=8192, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: is_llama_config=True, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: max_position_embeddings=32768, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: num_attention_heads=32, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: num_hidden_layers=16, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: num_key_value_heads=8, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: pad_token_id=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: pretraining_tp=1, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: rms_norm_eps=1e-05, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: rope_scaling=None, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: rope_theta=10000.0, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: rope_interleaved=False, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: tie_word_embeddings=True, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: use_cache=True, +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: vocab_size=32768) +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Building model.. +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Initialize RoPE Theta = 10000.0 +12/10/2024 03:05:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Setting PP block ranks... +12/10/2024 03:05:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Total number of parameters: 1.04G (1984.52MiB) +12/10/2024 03:05:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Local number of parameters: 260M (496.13MiB) +12/10/2024 03:05:02 [INFO|DP=0|PP=0|TP=2|ip-26-0-164-236]: Local number of parameters: 260M (496.13MiB) +12/10/2024 03:05:02 [INFO|DP=0|PP=0|TP=1|ip-26-0-164-236]: Local number of parameters: 260M (496.13MiB) +12/10/2024 03:05:02 [INFO|DP=0|PP=0|TP=3|ip-26-0-164-236]: Local number of parameters: 260M (496.13MiB) +12/10/2024 03:05:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: [After model building] Memory usage: 497.15MiB. Peak allocated: 5504.00MiB Peak reserved: 21890.00MiB +12/10/2024 03:05:02 [INFO|DP=0|PP=0|TP=2|ip-26-0-164-236]: [After model building] Memory usage: 497.15MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/10/2024 03:05:02 [INFO|DP=0|PP=0|TP=1|ip-26-0-164-236]: [After model building] Memory usage: 497.15MiB. Peak allocated: 5504.00MiB Peak reserved: 22914.00MiB +12/10/2024 03:05:02 [INFO|DP=0|PP=0|TP=3|ip-26-0-164-236]: [After model building] Memory usage: 497.15MiB. Peak allocated: 5504.00MiB Peak reserved: 22914.00MiB +12/10/2024 03:05:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: No checkpoint path provided. +12/10/2024 03:05:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Parametrizing model parameters using StandardParametrizator +12/10/2024 03:05:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: [Optimizer Building] Using LearningRateForSP as learning rate +12/10/2024 03:05:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: [ZeRO sharding] Size of optimizer params per rank: +12/10/2024 03:05:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: [ZeRO sharding] DP Rank 0 has 65M out of 260M (25.00%) params' optimizer states +12/10/2024 03:05:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: [ZeRO sharding] DP Rank 1 has 65M out of 260M (25.00%) params' optimizer states +12/10/2024 03:05:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: [ZeRO sharding] DP Rank 2 has 65M out of 260M (25.00%) params' optimizer states +12/10/2024 03:05:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: [ZeRO sharding] DP Rank 3 has 65M out of 260M (25.00%) params' optimizer states +12/10/2024 03:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/10/2024 03:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Using dummy data generator +12/10/2024 03:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: [Training Plan] There are 1 training stages +12/10/2024 03:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: [Stage Stable Training Stage] start from step 1 +12/10/2024 03:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: +12/10/2024 03:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: [Start training] datetime: 2024-12-10 03:05:05.434704 | mbs: 1 | grad_accum: 32 | global_batch_size: 128 | sequence_length: 32768 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/10/2024 03:05:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/10/2024 03:05:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Memory usage: 1737.48MiB. Peak allocated 5504.00MiB. Peak reserved: 21890.00MiB +12/10/2024 03:05:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Memory usage: 1866.99MiB. Peak allocated 21856.57MiB. Peak reserved: 28664.00MiB +12/10/2024 03:05:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 14.6K | tokens_per_sec: 288K | tokens_per_sec_per_gpu: 18K | global_batch_size: 128 | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 344 | hardware_tflops_per_gpu: 344 | grad_norm: 0.304 | cuda_memory_allocated: 2.49G | cuda_max_memory_reserved: 30.1G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.5G | hd_free_memory_tb: 244G +12/10/2024 03:05:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Memory usage: 2371.91MiB. Peak allocated 3488.20MiB. Peak reserved: 28674.00MiB +12/10/2024 03:05:40 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Memory usage: 2371.92MiB. Peak allocated 22361.51MiB. Peak reserved: 28696.00MiB +12/10/2024 03:05:40 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 12.4K | tokens_per_sec: 339K | tokens_per_sec_per_gpu: 21.2K | global_batch_size: 128 | lm_loss: 10.8 | lr: 0.0003 | model_tflops_per_gpu: 406 | hardware_tflops_per_gpu: 406 | grad_norm: 0.304 | cuda_memory_allocated: 2.49G | cuda_max_memory_reserved: 30.1G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.5G | hd_free_memory_tb: 244G +12/10/2024 03:05:40 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Memory usage: 2371.91MiB. Peak allocated 3488.22MiB. Peak reserved: 28696.00MiB +12/10/2024 03:05:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: Memory usage: 2371.92MiB. Peak allocated 22361.51MiB. Peak reserved: 28696.00MiB +num_params +num_params{'total': 1040457728, 'local': 260114432}num_paramsnum_params +num_params + + + +{'total': 1040457728, 'local': 260114432} +{'total': 1040457728, 'local': 260114432}{'total': 1040457728, 'local': 260114432} +num_params{'total': 1040457728, 'local': 260114432}num_params + + + +{'total': 1040457728, 'local': 260114432} +{'total': 1040457728, 'local': 260114432} +num_paramsnum_paramsnum_params + +num_params +{'total': 1040457728, 'local': 260114432}num_params{'total': 1040457728, 'local': 260114432} +num_params + + +{'total': 1040457728, 'local': 260114432} + +{'total': 1040457728, 'local': 260114432}{'total': 1040457728, 'local': 260114432}{'total': 1040457728, 'local': 260114432} + + +num_params +num_params{'total': 1040457728, 'local': 260114432} + +{'total': 1040457728, 'local': 260114432} +12/10/2024 03:05:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: iteration: 3 / 100 | consumed_tokens: 12.6M | elapsed_time_per_iteration_ms: 12.4K | tokens_per_sec: 338K | tokens_per_sec_per_gpu: 21.1K | global_batch_size: 128 | lm_loss: 10.8 | lr: 0.000296 | model_tflops_per_gpu: 404 | hardware_tflops_per_gpu: 404 | grad_norm: 0.295 | cuda_memory_allocated: 2.49G | cuda_max_memory_reserved: 30.1G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.5G | hd_free_memory_tb: 244G +num_params +{'total': 1040457728, 'local': 260114432} +12/10/2024 03:05:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/10/2024 03:05:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-164-236]: | -------- | -------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | ----------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/10/2024 03:05:53 [INFO|DP=2|PP=0|TP=2|ip-26-0-165-38]: Throughput logging complete +12/10/2024 03:05:53 [INFO|DP=3|PP=0|TP=0|ip-26-0-165-38]: Throughput logging complete +12/10/2024 03:05:53 [INFO|DP=2|PP=0|TP=1|ip-26-0-165-38]: Throughput logging complete +12/10/2024 03:05:53 [INFO|DP=2|PP=0|TP=3|ip-26-0-165-38]: Throughput logging complete +12/10/2024 03:05:53 [INFO|DP=3|PP=0|TP=1|ip-26-0-165-38]: Throughput logging complete +12/10/2024 03:05:53 [INFO|DP=3|PP=0|TP=3|ip-26-0-165-38]: Throughput logging complete +12/10/2024 03:05:53 [INFO|DP=2|PP=0|TP=0|ip-26-0-165-38]: Throughput logging complete +12/10/2024 03:05:53 [INFO|DP=3|PP=0|TP=2|ip-26-0-165-38]: Throughput logging complete +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-10 03:05:53,164] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 03:05:53,164] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 735821 closing signal SIGTERM +[2024-12-10 03:05:53,164] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 735822 closing signal SIGTERM +[2024-12-10 03:05:53,164] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 735823 closing signal SIGTERM +slurmstepd: error: *** JOB 13265609 ON ip-26-0-164-236 CANCELLED AT 2024-12-10T03:05:53 *** +[2024-12-10 03:05:53,164] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 735824 closing signal SIGTERM +[2024-12-10 03:05:53,164] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 735825 closing signal SIGTERM +[2024-12-10 03:05:53,166] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 735826 closing signal SIGTERM +[2024-12-10 03:05:53,167] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 735827 closing signal SIGTERM +[2024-12-10 03:05:53,167] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 735828 closing signal SIGTERM diff --git a/logs/13265614-bench_1.34G_dp4_tp4_pp1_acc4_mbs2_seq32768_zero1_tpmodeRED_vocab131k.out b/logs/13265614-bench_1.34G_dp4_tp4_pp1_acc4_mbs2_seq32768_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..ee5d4d6fcd7224338e1b8bc55f484e708c8ccc58 --- /dev/null +++ b/logs/13265614-bench_1.34G_dp4_tp4_pp1_acc4_mbs2_seq32768_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,657 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-166-15,ip-26-0-169-132 ++ export 'NODELIST=ip-26-0-166-15 +ip-26-0-169-132' ++ NODELIST='ip-26-0-166-15 +ip-26-0-169-132' +++ scontrol show hostnames ip-26-0-166-15,ip-26-0-169-132 +++ head -n1 ++ export MASTER_NODE=ip-26-0-166-15 ++ MASTER_NODE=ip-26-0-166-15 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-166-15' +Master node: ip-26-0-166-15 ++ echo 'All nodes: ip-26-0-166-15 +ip-26-0-169-132' +All nodes: ip-26-0-166-15 +ip-26-0-169-132 ++ echo 'World size: 16' +World size: 16 ++ srun torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13265614 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-166-15:12356 run_train.py --config-file benchmark/configs/config_1.34G_dp4_tp4_pp1_acc4_mbs2_seq32768_zero1_tpmodeRED_vocab131k.yaml +[2024-12-10 03:06:37,844] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 03:06:37,855] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 03:06:37,844] torch.distributed.run: [WARNING] +[2024-12-10 03:06:37,844] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 03:06:37,844] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 03:06:37,844] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 03:06:37,855] torch.distributed.run: [WARNING] +[2024-12-10 03:06:37,855] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 03:06:37,855] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 03:06:37,855] torch.distributed.run: [WARNING] ***************************************** +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Config: +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Config(general=GeneralArgs(project='debug', +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: run='1.34G_dp4_tp4_pp1_acc4_mbs2_seq32768_zero1_tpmodeRED_vocab131k', +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: seed=42, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: step=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: consumed_train_samples=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: ignore_sanity_checks=True), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: parallelism=ParallelismArgs(dp=4, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pp=1, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tp=4, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pp_engine=, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tp_mode=, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tp_linear_async_communication=True, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: recompute_layer=False, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tp_recompute_allgather=True, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: expert_parallel_size=1), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: eos_token_id=0, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: hidden_act='silu', +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: hidden_size=2048, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: initializer_range=0.02, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: intermediate_size=8192, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: is_llama_config=True, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: max_position_embeddings=32768, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_attention_heads=32, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_hidden_layers=16, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_key_value_heads=8, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pad_token_id=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pretraining_tp=1, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rms_norm_eps=1e-05, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_scaling=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_theta=10000.0, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_interleaved=False, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tie_word_embeddings=True, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: use_cache=True, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: vocab_size=131072), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: init_method=RandomInit(std=0.02), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: dtype=torch.bfloat16, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: make_vocab_size_divisible_by=1, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: ddp_bucket_cap_mb=25), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tokenizer_revision=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tokenizer_max_length=None), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: checkpoint_interval=10000, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: save_initial_state=False, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: save_final_state=False, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: resume_checkpoint_path=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: checkpoints_path_is_shared_file_system=False), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: logging=LoggingArgs(log_level='info', +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: log_level_replica='info', +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: iteration_step_info_interval=1), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tokens=TokensArgs(sequence_length=32768, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: train_steps=100, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: micro_batch_size=2, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: batch_accumulation_per_replica=4, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: val_check_interval=100, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: limit_val_batches=0, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: limit_test_batches=0), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: adam_beta1=0.9, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: adam_beta2=0.95, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: torch_adam_is_fused=True, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: name='adamW'), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: zero_stage=1, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: weight_decay=0.01, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: clip_grad=1.0, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: accumulate_grad_in_fp32=True, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lr_warmup_steps=2, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lr_warmup_style='linear', +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lr_decay_style='cosine', +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lr_decay_steps=13, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lr_decay_starting_step=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: min_decay_lr=1e-05)), +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: start_training_step=1, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: data=DataArgs(dataset=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: seed=42, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_loading_workers=1))], +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: profiler=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: lighteval=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: s3_upload=None) +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Model Config: +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: LlamaConfig(bos_token_id=0, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: eos_token_id=0, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: hidden_act='silu', +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: hidden_size=2048, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: initializer_range=0.02, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: intermediate_size=8192, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: is_llama_config=True, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: max_position_embeddings=32768, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_attention_heads=32, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_hidden_layers=16, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: num_key_value_heads=8, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pad_token_id=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: pretraining_tp=1, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rms_norm_eps=1e-05, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_scaling=None, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_theta=10000.0, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: rope_interleaved=False, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: tie_word_embeddings=True, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: use_cache=True, +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: vocab_size=131072) +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Building model.. +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Initialize RoPE Theta = 10000.0 +12/10/2024 03:07:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Setting PP block ranks... +12/10/2024 03:07:21 [INFO|DP=0|PP=0|TP=3|ip-26-0-166-15]: Local number of parameters: 310M (592.13MiB) +12/10/2024 03:07:21 [INFO|DP=0|PP=0|TP=1|ip-26-0-166-15]: Local number of parameters: 310M (592.13MiB) +12/10/2024 03:07:21 [INFO|DP=0|PP=0|TP=2|ip-26-0-166-15]: Local number of parameters: 310M (592.13MiB) +12/10/2024 03:07:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Total number of parameters: 1.24G (2368.52MiB) +12/10/2024 03:07:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Local number of parameters: 310M (592.13MiB) +12/10/2024 03:07:21 [INFO|DP=0|PP=0|TP=3|ip-26-0-166-15]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 03:07:21 [INFO|DP=0|PP=0|TP=2|ip-26-0-166-15]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 03:07:21 [INFO|DP=0|PP=0|TP=1|ip-26-0-166-15]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 03:07:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [After model building] Memory usage: 595.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35202.00MiB +12/10/2024 03:07:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: No checkpoint path provided. +12/10/2024 03:07:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Parametrizing model parameters using StandardParametrizator +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [Optimizer Building] Using LearningRateForSP as learning rate +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [ZeRO sharding] Size of optimizer params per rank: +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [ZeRO sharding] DP Rank 0 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [ZeRO sharding] DP Rank 1 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [ZeRO sharding] DP Rank 2 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [ZeRO sharding] DP Rank 3 has 77.6M out of 310M (25.00%) params' optimizer states +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Using dummy data generator +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [Training Plan] There are 1 training stages +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [Stage Stable Training Stage] start from step 1 +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: +12/10/2024 03:07:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: [Start training] datetime: 2024-12-10 03:07:23.986233 | mbs: 2 | grad_accum: 4 | global_batch_size: 32 | sequence_length: 32768 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/10/2024 03:07:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/10/2024 03:07:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Memory usage: 2075.48MiB. Peak allocated 5504.00MiB. Peak reserved: 36388.00MiB +12/10/2024 03:07:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Memory usage: 2716.60MiB. Peak allocated 42306.06MiB. Peak reserved: 57194.00MiB +12/10/2024 03:07:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 5.19K | tokens_per_sec: 202K | tokens_per_sec_per_gpu: 12.6K | global_batch_size: 32 | lm_loss: 12.2 | lr: 0.00015 | model_tflops_per_gpu: 257 | hardware_tflops_per_gpu: 257 | grad_norm: 0.6 | cuda_memory_allocated: 3.48G | cuda_max_memory_reserved: 60G | hd_total_memory_tb: 312G | hd_used_memory_tb: 71G | hd_free_memory_tb: 241G +12/10/2024 03:07:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Memory usage: 3319.46MiB. Peak allocated 4651.75MiB. Peak reserved: 57220.00MiB +12/10/2024 03:07:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Memory usage: 3319.46MiB. Peak allocated 42908.92MiB. Peak reserved: 57226.00MiB +12/10/2024 03:07:40 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 3.27K | tokens_per_sec: 321K | tokens_per_sec_per_gpu: 20K | global_batch_size: 32 | lm_loss: 12.2 | lr: 0.0003 | model_tflops_per_gpu: 407 | hardware_tflops_per_gpu: 407 | grad_norm: 0.6 | cuda_memory_allocated: 3.48G | cuda_max_memory_reserved: 60G | hd_total_memory_tb: 312G | hd_used_memory_tb: 71G | hd_free_memory_tb: 241G +12/10/2024 03:07:40 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Memory usage: 3319.46MiB. Peak allocated 4651.75MiB. Peak reserved: 57226.00MiB +12/10/2024 03:07:43 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Memory usage: 3319.46MiB. Peak allocated 42908.92MiB. Peak reserved: 57226.00MiB +num_paramsnum_params +num_params +{'total': 1241784320, 'local': 310446080} + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +num_params +num_paramsnum_params{'total': 1241784320, 'local': 310446080} + + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +num_paramsnum_params + +{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080} + +num_params +num_params +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +num_paramsnum_params +num_params +{'total': 1241784320, 'local': 310446080}num_params + +{'total': 1241784320, 'local': 310446080} + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +num_params +{'total': 1241784320, 'local': 310446080} +12/10/2024 03:07:43 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 3.28K | tokens_per_sec: 320K | tokens_per_sec_per_gpu: 20K | global_batch_size: 32 | lm_loss: 12.2 | lr: 0.000296 | model_tflops_per_gpu: 407 | hardware_tflops_per_gpu: 407 | grad_norm: 0.589 | cuda_memory_allocated: 3.48G | cuda_max_memory_reserved: 60G | hd_total_memory_tb: 312G | hd_used_memory_tb: 71G | hd_free_memory_tb: 241G +num_params +{'total': 1241784320, 'local': 310446080} +12/10/2024 03:07:43 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/10/2024 03:07:43 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: | -------- | -------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | --------------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/10/2024 03:07:43 [INFO|DP=2|PP=0|TP=1|ip-26-0-169-132]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=2|PP=0|TP=0|ip-26-0-169-132]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=2|PP=0|TP=2|ip-26-0-169-132]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=3|PP=0|TP=2|ip-26-0-169-132]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=3|PP=0|TP=0|ip-26-0-169-132]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=2|PP=0|TP=3|ip-26-0-169-132]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=3|PP=0|TP=1|ip-26-0-169-132]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=3|PP=0|TP=3|ip-26-0-169-132]: Throughput logging complete +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-10 03:07:43,660] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +slurmstepd: error: *** JOB 13265614 ON ip-26-0-166-15 CANCELLED AT 2024-12-10T03:07:43 *** +[2024-12-10 03:07:43,661] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 640363 closing signal SIGTERM +[2024-12-10 03:07:43,661] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 640364 closing signal SIGTERM +[2024-12-10 03:07:43,661] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 640365 closing signal SIGTERM +[2024-12-10 03:07:43,661] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 640366 closing signal SIGTERM +[2024-12-10 03:07:43,663] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 640367 closing signal SIGTERM +[2024-12-10 03:07:43,663] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 640368 closing signal SIGTERM +[2024-12-10 03:07:43,664] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 640369 closing signal SIGTERM +[2024-12-10 03:07:43,664] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 640370 closing signal SIGTERM +12/10/2024 03:07:43 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: | 13265614 | 1.34G_dp4_tp4_pp1_acc4_mbs2_seq32768_zero1_tpmodeRED_vocab131k | 2 | 32768 | 2 | 4 | 32 | 406.88 | 406.88 | 20009.71 | 436.06 | 222.62 | 224.98 | 461.32 | 264.57 | 263.37 | 4.54 | 55.88 | 4 | 1 | 4 | | TensorParallelLinearMode.REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 8 | 32768 | 131072 | True | torch.bfloat16 | 1 | 25 | True | 1.24G | 310M | +12/10/2024 03:07:43 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=0|PP=0|TP=1|ip-26-0-166-15]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=0|PP=0|TP=2|ip-26-0-166-15]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=0|PP=0|TP=3|ip-26-0-166-15]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=1|PP=0|TP=1|ip-26-0-166-15]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=1|PP=0|TP=3|ip-26-0-166-15]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=1|PP=0|TP=0|ip-26-0-166-15]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=1|PP=0|TP=2|ip-26-0-166-15]: Throughput logging complete +12/10/2024 03:07:43 [INFO|DP=0|PP=0|TP=0|ip-26-0-166-15]: Memory usage: 3319.46MiB. Peak allocated 4651.75MiB. Peak reserved: 57226.00MiB +slurmstepd: error: *** STEP 13265614.0 ON ip-26-0-166-15 CANCELLED AT 2024-12-10T03:07:43 *** +[2024-12-10 03:07:43,660] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 03:07:43,661] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68850 closing signal SIGTERM +[2024-12-10 03:07:43,661] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68851 closing signal SIGTERM +[2024-12-10 03:07:43,661] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68852 closing signal SIGTERM +[2024-12-10 03:07:43,661] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68853 closing signal SIGTERM +[2024-12-10 03:07:43,662] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68854 closing signal SIGTERM +[2024-12-10 03:07:43,662] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68855 closing signal SIGTERM +[2024-12-10 03:07:43,662] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68856 closing signal SIGTERM +[2024-12-10 03:07:43,663] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68857 closing signal SIGTERM diff --git a/logs/13265718-bench_1.34G_dp8_tp4_pp1_acc8_mbs2_seq8192_zero1_tpmodeRED_vocab131k.out b/logs/13265718-bench_1.34G_dp8_tp4_pp1_acc8_mbs2_seq8192_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..0219c4df7f2ffec93ab2bd298021049d98aed8e1 --- /dev/null +++ b/logs/13265718-bench_1.34G_dp8_tp4_pp1_acc8_mbs2_seq8192_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,743 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-170-143,ip-26-0-171-[21,230],ip-26-0-175-241' ++ export 'NODELIST=ip-26-0-170-143 +ip-26-0-171-21 +ip-26-0-171-230 +ip-26-0-175-241' ++ NODELIST='ip-26-0-170-143 +ip-26-0-171-21 +ip-26-0-171-230 +ip-26-0-175-241' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-170-143,ip-26-0-171-[21,230],ip-26-0-175-241' ++ export MASTER_NODE=ip-26-0-170-143 ++ MASTER_NODE=ip-26-0-170-143 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-170-143' +Master node: ip-26-0-170-143 ++ echo 'All nodes: ip-26-0-170-143 +ip-26-0-171-21 +ip-26-0-171-230 +ip-26-0-175-241' +All nodes: ip-26-0-170-143 +ip-26-0-171-21 +ip-26-0-171-230 +ip-26-0-175-241 ++ echo 'World size: 32' +World size: 32 ++ srun torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13265718 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-170-143:12356 run_train.py --config-file benchmark/configs/config_1.34G_dp8_tp4_pp1_acc8_mbs2_seq8192_zero1_tpmodeRED_vocab131k.yaml +[2024-12-10 05:36:09,330] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 05:36:09,330] torch.distributed.run: [WARNING] +[2024-12-10 05:36:09,330] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 05:36:09,330] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 05:36:09,330] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 05:36:09,434] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 05:36:09,434] torch.distributed.run: [WARNING] +[2024-12-10 05:36:09,434] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 05:36:09,434] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 05:36:09,434] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 05:36:09,626] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 05:36:09,626] torch.distributed.run: [WARNING] +[2024-12-10 05:36:09,626] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 05:36:09,626] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 05:36:09,626] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 05:36:14,748] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 05:36:14,749] torch.distributed.run: [WARNING] +[2024-12-10 05:36:14,749] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 05:36:14,749] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 05:36:14,749] torch.distributed.run: [WARNING] ***************************************** +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Config: +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Config(general=GeneralArgs(project='debug', +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: run='1.34G_dp8_tp4_pp1_acc8_mbs2_seq8192_zero1_tpmodeRED_vocab131k', +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: seed=42, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: step=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: consumed_train_samples=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: ignore_sanity_checks=True), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: parallelism=ParallelismArgs(dp=8, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pp=1, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp=4, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pp_engine=, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp_mode=, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp_linear_async_communication=True, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: recompute_layer=False, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp_recompute_allgather=True, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: expert_parallel_size=1), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: eos_token_id=0, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_act='silu', +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_size=2048, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: initializer_range=0.02, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: intermediate_size=8192, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: is_llama_config=True, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: max_position_embeddings=8192, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_attention_heads=32, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_hidden_layers=16, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_key_value_heads=8, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pad_token_id=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pretraining_tp=1, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rms_norm_eps=1e-05, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_scaling=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_theta=10000.0, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_interleaved=False, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tie_word_embeddings=True, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: use_cache=True, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: vocab_size=131072), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: init_method=RandomInit(std=0.02), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: dtype=torch.bfloat16, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: make_vocab_size_divisible_by=1, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: ddp_bucket_cap_mb=25), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokenizer_revision=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokenizer_max_length=None), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: checkpoint_interval=10000, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: save_initial_state=False, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: save_final_state=False, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: resume_checkpoint_path=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: checkpoints_path_is_shared_file_system=False), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: logging=LoggingArgs(log_level='info', +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: log_level_replica='info', +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: iteration_step_info_interval=1), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokens=TokensArgs(sequence_length=8192, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: train_steps=100, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: micro_batch_size=2, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: batch_accumulation_per_replica=8, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: val_check_interval=100, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: limit_val_batches=0, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: limit_test_batches=0), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: adam_beta1=0.9, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: adam_beta2=0.95, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: torch_adam_is_fused=True, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: name='adamW'), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: zero_stage=1, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: weight_decay=0.01, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: clip_grad=1.0, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: accumulate_grad_in_fp32=True, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_warmup_steps=2, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_warmup_style='linear', +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_decay_style='cosine', +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_decay_steps=13, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_decay_starting_step=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: min_decay_lr=1e-05)), +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: start_training_step=1, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: data=DataArgs(dataset=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: seed=42, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_loading_workers=1))], +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: profiler=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lighteval=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: s3_upload=None) +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Model Config: +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: LlamaConfig(bos_token_id=0, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: eos_token_id=0, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_act='silu', +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_size=2048, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: initializer_range=0.02, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: intermediate_size=8192, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: is_llama_config=True, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: max_position_embeddings=8192, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_attention_heads=32, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_hidden_layers=16, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_key_value_heads=8, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pad_token_id=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pretraining_tp=1, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rms_norm_eps=1e-05, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_scaling=None, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_theta=10000.0, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_interleaved=False, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tie_word_embeddings=True, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: use_cache=True, +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: vocab_size=131072) +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Building model.. +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Initialize RoPE Theta = 10000.0 +12/10/2024 05:36:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Setting PP block ranks... +12/10/2024 05:37:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Total number of parameters: 1.24G (2368.52MiB) +12/10/2024 05:37:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Local number of parameters: 310M (592.13MiB) +12/10/2024 05:37:03 [INFO|DP=0|PP=0|TP=2|ip-26-0-170-143]: Local number of parameters: 310M (592.13MiB) +12/10/2024 05:37:03 [INFO|DP=0|PP=0|TP=3|ip-26-0-170-143]: Local number of parameters: 310M (592.13MiB) +12/10/2024 05:37:03 [INFO|DP=0|PP=0|TP=1|ip-26-0-170-143]: Local number of parameters: 310M (592.13MiB) +12/10/2024 05:37:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [After model building] Memory usage: 594.15MiB. Peak allocated: 5440.00MiB Peak reserved: 24898.00MiB +12/10/2024 05:37:03 [INFO|DP=0|PP=0|TP=2|ip-26-0-170-143]: [After model building] Memory usage: 594.15MiB. Peak allocated: 5440.00MiB Peak reserved: 26946.00MiB +12/10/2024 05:37:03 [INFO|DP=0|PP=0|TP=3|ip-26-0-170-143]: [After model building] Memory usage: 594.15MiB. Peak allocated: 5440.00MiB Peak reserved: 24898.00MiB +12/10/2024 05:37:03 [INFO|DP=0|PP=0|TP=1|ip-26-0-170-143]: [After model building] Memory usage: 594.15MiB. Peak allocated: 5440.00MiB Peak reserved: 26946.00MiB +12/10/2024 05:37:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: No checkpoint path provided. +12/10/2024 05:37:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Parametrizing model parameters using StandardParametrizator +12/10/2024 05:37:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Optimizer Building] Using LearningRateForSP as learning rate +12/10/2024 05:37:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] Size of optimizer params per rank: +12/10/2024 05:37:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] DP Rank 0 has 38.8M out of 310M (12.50%) params' optimizer states +12/10/2024 05:37:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] DP Rank 1 has 38.8M out of 310M (12.50%) params' optimizer states +12/10/2024 05:37:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] DP Rank 2 has 38.8M out of 310M (12.50%) params' optimizer states +12/10/2024 05:37:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] DP Rank 3 has 38.8M out of 310M (12.50%) params' optimizer states +12/10/2024 05:37:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] DP Rank 4 has 38.8M out of 310M (12.50%) params' optimizer states +12/10/2024 05:37:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] DP Rank 5 has 38.8M out of 310M (12.50%) params' optimizer states +12/10/2024 05:37:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] DP Rank 6 has 38.8M out of 310M (12.50%) params' optimizer states +12/10/2024 05:37:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] DP Rank 7 has 38.8M out of 310M (12.50%) params' optimizer states +12/10/2024 05:37:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/10/2024 05:37:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Using dummy data generator +12/10/2024 05:37:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Training Plan] There are 1 training stages +12/10/2024 05:37:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Stage Stable Training Stage] start from step 1 +12/10/2024 05:37:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: +12/10/2024 05:37:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Start training] datetime: 2024-12-10 05:37:07.493459 | mbs: 2 | grad_accum: 8 | global_batch_size: 128 | sequence_length: 8192 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/10/2024 05:37:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/10/2024 05:37:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 1926.44MiB. Peak allocated 5440.00MiB. Peak reserved: 26084.00MiB +12/10/2024 05:37:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 2134.73MiB. Peak allocated 12032.11MiB. Peak reserved: 17190.00MiB +12/10/2024 05:37:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 3.22K | tokens_per_sec: 326K | tokens_per_sec_per_gpu: 10.2K | global_batch_size: 128 | lm_loss: 12.2 | lr: 0.00015 | model_tflops_per_gpu: 109 | hardware_tflops_per_gpu: 109 | grad_norm: 0.618 | cuda_memory_allocated: 2.55G | cuda_max_memory_reserved: 18.1G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.2G | hd_free_memory_tb: 242G +12/10/2024 05:37:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 2430.84MiB. Peak allocated 3689.12MiB. Peak reserved: 17222.00MiB +12/10/2024 05:37:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 2430.85MiB. Peak allocated 12328.23MiB. Peak reserved: 17230.00MiB +12/10/2024 05:37:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 1.21K | tokens_per_sec: 864K | tokens_per_sec_per_gpu: 27K | global_batch_size: 128 | lm_loss: 12.2 | lr: 0.0003 | model_tflops_per_gpu: 288 | hardware_tflops_per_gpu: 288 | grad_norm: 0.619 | cuda_memory_allocated: 2.55G | cuda_max_memory_reserved: 18.1G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.2G | hd_free_memory_tb: 242G +12/10/2024 05:37:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 2430.84MiB. Peak allocated 3689.12MiB. Peak reserved: 17230.00MiB +12/10/2024 05:37:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 2430.85MiB. Peak allocated 12328.23MiB. Peak reserved: 17230.00MiB +num_paramsnum_params + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +num_params +num_params{'total': 1241784320, 'local': 310446080} + +num_params{'total': 1241784320, 'local': 310446080} + +num_params +num_paramsnum_params{'total': 1241784320, 'local': 310446080} + + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +num_paramsnum_paramsnum_paramsnum_params + +num_params +num_params{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} + + +{'total': 1241784320, 'local': 310446080} + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080}num_params + +num_params +{'total': 1241784320, 'local': 310446080} +num_params +{'total': 1241784320, 'local': 310446080} +num_paramsnum_paramsnum_params + + +num_params{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080}num_params + + + +num_params{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080} + + +num_params{'total': 1241784320, 'local': 310446080} + +num_params{'total': 1241784320, 'local': 310446080} + +{'total': 1241784320, 'local': 310446080} +num_paramsnum_params + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +num_params +{'total': 1241784320, 'local': 310446080} +num_params +{'total': 1241784320, 'local': 310446080} +num_params +{'total': 1241784320, 'local': 310446080} +num_params +{'total': 1241784320, 'local': 310446080} +12/10/2024 05:37:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 1.11K | tokens_per_sec: 942K | tokens_per_sec_per_gpu: 29.4K | global_batch_size: 128 | lm_loss: 12.2 | lr: 0.000296 | model_tflops_per_gpu: 314 | hardware_tflops_per_gpu: 314 | grad_norm: 0.611 | cuda_memory_allocated: 2.55G | cuda_max_memory_reserved: 18.1G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.2G | hd_free_memory_tb: 242G +num_params +{'total': 1241784320, 'local': 310446080} +12/10/2024 05:37:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/10/2024 05:37:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: | -------- | ------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | --------------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/10/2024 05:37:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: | 13265718 | 1.34G_dp8_tp4_pp1_acc8_mbs2_seq8192_zero1_tpmodeRED_vocab131k | 4 | 8192 | 2 | 8 | 128 | 314.01 | 314.01 | 29428.84 | 244.95 | 208.89 | 209.97 | 461.24 | 265.10 | 265.39 | 3.60 | 16.83 | 8 | 1 | 4 | | TensorParallelLinearMode.REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 8 | 8192 | 131072 | True | torch.bfloat16 | 1 | 25 | True | 1.24G | 310M | +12/10/2024 05:37:22 [INFO|DP=4|PP=0|TP=3|ip-26-0-171-230]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=4|PP=0|TP=1|ip-26-0-171-230]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=5|PP=0|TP=0|ip-26-0-171-230]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=4|PP=0|TP=0|ip-26-0-171-230]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=5|PP=0|TP=1|ip-26-0-171-230]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=2|PP=0|TP=2|ip-26-0-171-21]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=2|PP=0|TP=0|ip-26-0-171-21]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=3|PP=0|TP=0|ip-26-0-171-21]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=5|PP=0|TP=2|ip-26-0-171-230]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=5|PP=0|TP=3|ip-26-0-171-230]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=4|PP=0|TP=2|ip-26-0-171-230]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=2|PP=0|TP=3|ip-26-0-171-21]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=2|PP=0|TP=1|ip-26-0-171-21]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=3|PP=0|TP=1|ip-26-0-171-21]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=3|PP=0|TP=2|ip-26-0-171-21]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=7|PP=0|TP=0|ip-26-0-175-241]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=7|PP=0|TP=1|ip-26-0-175-241]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=6|PP=0|TP=0|ip-26-0-175-241]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=7|PP=0|TP=2|ip-26-0-175-241]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=3|PP=0|TP=3|ip-26-0-171-21]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=6|PP=0|TP=1|ip-26-0-175-241]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=7|PP=0|TP=3|ip-26-0-175-241]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=6|PP=0|TP=3|ip-26-0-175-241]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=6|PP=0|TP=2|ip-26-0-175-241]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=0|PP=0|TP=2|ip-26-0-170-143]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=0|PP=0|TP=1|ip-26-0-170-143]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=0|PP=0|TP=3|ip-26-0-170-143]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=1|PP=0|TP=3|ip-26-0-170-143]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=1|PP=0|TP=1|ip-26-0-170-143]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=1|PP=0|TP=2|ip-26-0-170-143]: Throughput logging complete +12/10/2024 05:37:22 [INFO|DP=1|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-10 05:37:22,218] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 05:37:22,218] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2989371 closing signal SIGTERM +[2024-12-10 05:37:22,218] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2989372 closing signal SIGTERM +[2024-12-10 05:37:22,219] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2989373 closing signal SIGTERM +[2024-12-10 05:37:22,219] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +12/10/2024 05:37:22 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 2430.84MiB. Peak allocated 3689.12MiB. Peak reserved: 17230.00MiB +[2024-12-10 05:37:22,219] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2989374 closing signal SIGTERM +[2024-12-10 05:37:22,219] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 115807 closing signal SIGTERM +[2024-12-10 05:37:22,219] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 115808 closing signal SIGTERM +[2024-12-10 05:37:22,219] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 115809 closing signal SIGTERM +slurmstepd: error: *** JOB 13265718 ON ip-26-0-170-143 CANCELLED AT 2024-12-10T05:37:22 *** +[2024-12-10 05:37:22,219] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 115810 closing signal SIGTERM +[2024-12-10 05:37:22,220] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 05:37:22,221] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3327482 closing signal SIGTERM +[2024-12-10 05:37:22,221] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3327483 closing signal SIGTERM +[2024-12-10 05:37:22,221] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3327484 closing signal SIGTERM +[2024-12-10 05:37:22,221] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3327485 closing signal SIGTERM +slurmstepd: error: *** STEP 13265718.0 ON ip-26-0-170-143 CANCELLED AT 2024-12-10T05:37:22 *** +[2024-12-10 05:37:22,221] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 115811 closing signal SIGTERM +[2024-12-10 05:37:22,221] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2989375 closing signal SIGTERM +[2024-12-10 05:37:22,221] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 115812 closing signal SIGTERM +[2024-12-10 05:37:22,221] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 115813 closing signal SIGTERM +[2024-12-10 05:37:22,222] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 05:37:22,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1700447 closing signal SIGTERM +[2024-12-10 05:37:22,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1700448 closing signal SIGTERM +[2024-12-10 05:37:22,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1700449 closing signal SIGTERM +[2024-12-10 05:37:22,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1700450 closing signal SIGTERM +[2024-12-10 05:37:22,221] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2989376 closing signal SIGTERM +[2024-12-10 05:37:22,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1700451 closing signal SIGTERM +[2024-12-10 05:37:22,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1700452 closing signal SIGTERM +[2024-12-10 05:37:22,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 115814 closing signal SIGTERM +[2024-12-10 05:37:22,223] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3327486 closing signal SIGTERM +[2024-12-10 05:37:22,221] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2989377 closing signal SIGTERM +[2024-12-10 05:37:22,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2989378 closing signal SIGTERM +[2024-12-10 05:37:22,223] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3327487 closing signal SIGTERM +[2024-12-10 05:37:22,223] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3327488 closing signal SIGTERM +[2024-12-10 05:37:22,223] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3327489 closing signal SIGTERM diff --git a/logs/13265818-bench_1.34G_dp16_tp4_pp1_acc16_mbs8_seq2048_zero1_tpmodeRED_vocab131k.out b/logs/13265818-bench_1.34G_dp16_tp4_pp1_acc16_mbs8_seq2048_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..c3f76df0b5bce72375c268bb9c439afe76187ef7 --- /dev/null +++ b/logs/13265818-bench_1.34G_dp16_tp4_pp1_acc16_mbs8_seq2048_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,920 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-160-225,ip-26-0-164-[45,75,187,207,236],ip-26-0-165-[24,38]' ++ export 'NODELIST=ip-26-0-160-225 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38' ++ NODELIST='ip-26-0-160-225 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-160-225,ip-26-0-164-[45,75,187,207,236],ip-26-0-165-[24,38]' ++ export MASTER_NODE=ip-26-0-160-225 ++ MASTER_NODE=ip-26-0-160-225 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=8 ++ NNODES=8 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=64 ++ WORLD_SIZE=64 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-225' +Master node: ip-26-0-160-225 ++ echo 'All nodes: ip-26-0-160-225 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38' +All nodes: ip-26-0-160-225 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 ++ echo 'World size: 64' +World size: 64 ++ srun torchrun --nnodes=8 --nproc_per_node=8 --rdzv_id=13265818 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-225:12356 run_train.py --config-file benchmark/configs/config_1.34G_dp16_tp4_pp1_acc16_mbs8_seq2048_zero1_tpmodeRED_vocab131k.yaml +[2024-12-10 06:49:59,616] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 06:49:59,620] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 06:49:59,622] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 06:49:59,623] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 06:49:59,627] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 06:49:59,627] torch.distributed.run: [WARNING] +[2024-12-10 06:49:59,627] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,627] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 06:49:59,627] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,649] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 06:49:59,616] torch.distributed.run: [WARNING] +[2024-12-10 06:49:59,616] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,616] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 06:49:59,616] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,621] torch.distributed.run: [WARNING] +[2024-12-10 06:49:59,621] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,621] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 06:49:59,621] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,623] torch.distributed.run: [WARNING] +[2024-12-10 06:49:59,623] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,623] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 06:49:59,623] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,622] torch.distributed.run: [WARNING] +[2024-12-10 06:49:59,622] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,622] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 06:49:59,622] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,649] torch.distributed.run: [WARNING] +[2024-12-10 06:49:59,649] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,649] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 06:49:59,649] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,707] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 06:49:59,708] torch.distributed.run: [WARNING] +[2024-12-10 06:49:59,708] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,708] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 06:49:59,708] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,785] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 06:49:59,785] torch.distributed.run: [WARNING] +[2024-12-10 06:49:59,785] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 06:49:59,785] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 06:49:59,785] torch.distributed.run: [WARNING] ***************************************** +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Config: +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Config(general=GeneralArgs(project='debug', +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: run='1.34G_dp16_tp4_pp1_acc16_mbs8_seq2048_zero1_tpmodeRED_vocab131k', +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: seed=42, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: step=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: consumed_train_samples=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: ignore_sanity_checks=True), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: parallelism=ParallelismArgs(dp=16, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pp=1, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp=4, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pp_engine=, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp_mode=, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp_linear_async_communication=True, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: recompute_layer=False, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp_recompute_allgather=True, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: expert_parallel_size=1), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: eos_token_id=0, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_act='silu', +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_size=2048, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: initializer_range=0.02, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: intermediate_size=8192, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: is_llama_config=True, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: max_position_embeddings=2048, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_attention_heads=32, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_hidden_layers=16, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_key_value_heads=8, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pad_token_id=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pretraining_tp=1, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rms_norm_eps=1e-05, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_scaling=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_theta=10000.0, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_interleaved=False, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tie_word_embeddings=True, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: use_cache=True, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: vocab_size=131072), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: init_method=RandomInit(std=0.02), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: dtype=torch.bfloat16, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: make_vocab_size_divisible_by=1, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: ddp_bucket_cap_mb=25), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer_revision=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer_max_length=None), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoint_interval=10000, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: save_initial_state=False, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: save_final_state=False, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: resume_checkpoint_path=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoints_path_is_shared_file_system=False), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: logging=LoggingArgs(log_level='info', +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: log_level_replica='info', +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: iteration_step_info_interval=1), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokens=TokensArgs(sequence_length=2048, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: train_steps=100, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: micro_batch_size=8, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: batch_accumulation_per_replica=16, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: val_check_interval=100, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: limit_val_batches=0, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: limit_test_batches=0), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: adam_beta1=0.9, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: adam_beta2=0.95, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: torch_adam_is_fused=True, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: name='adamW'), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: zero_stage=1, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: weight_decay=0.01, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: clip_grad=1.0, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: accumulate_grad_in_fp32=True, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_warmup_steps=2, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_warmup_style='linear', +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_style='cosine', +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_steps=13, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_starting_step=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: min_decay_lr=1e-05)), +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: start_training_step=1, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: data=DataArgs(dataset=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: seed=42, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_loading_workers=1))], +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: profiler=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lighteval=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: s3_upload=None) +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Model Config: +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: LlamaConfig(bos_token_id=0, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: eos_token_id=0, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_act='silu', +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_size=2048, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: initializer_range=0.02, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: intermediate_size=8192, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: is_llama_config=True, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: max_position_embeddings=2048, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_attention_heads=32, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_hidden_layers=16, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_key_value_heads=8, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pad_token_id=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pretraining_tp=1, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rms_norm_eps=1e-05, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_scaling=None, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_theta=10000.0, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_interleaved=False, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tie_word_embeddings=True, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: use_cache=True, +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: vocab_size=131072) +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Building model.. +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Initialize RoPE Theta = 10000.0 +12/10/2024 06:50:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Setting PP block ranks... +12/10/2024 06:50:46 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-225]: Local number of parameters: 310M (592.13MiB) +12/10/2024 06:50:46 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-225]: Local number of parameters: 310M (592.13MiB) +12/10/2024 06:50:46 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-225]: Local number of parameters: 310M (592.13MiB) +12/10/2024 06:50:46 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-225]: [After model building] Memory usage: 592.15MiB. Peak allocated: 5408.00MiB Peak reserved: 28962.00MiB +12/10/2024 06:50:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Total number of parameters: 1.24G (2368.52MiB) +12/10/2024 06:50:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Local number of parameters: 310M (592.13MiB) +12/10/2024 06:50:46 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-225]: [After model building] Memory usage: 592.15MiB. Peak allocated: 5408.00MiB Peak reserved: 29986.00MiB +12/10/2024 06:50:46 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-225]: [After model building] Memory usage: 592.15MiB. Peak allocated: 5408.00MiB Peak reserved: 31010.00MiB +12/10/2024 06:50:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [After model building] Memory usage: 592.15MiB. Peak allocated: 5408.00MiB Peak reserved: 29986.00MiB +12/10/2024 06:50:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: No checkpoint path provided. +12/10/2024 06:50:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Parametrizing model parameters using StandardParametrizator +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Optimizer Building] Using LearningRateForSP as learning rate +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] Size of optimizer params per rank: +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 0 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 1 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 2 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 3 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 4 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 5 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 6 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 7 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 8 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 9 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 10 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 11 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 12 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 13 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 14 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [ZeRO sharding] DP Rank 15 has 19.4M out of 310M (6.25%) params' optimizer states +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Using dummy data generator +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Training Plan] There are 1 training stages +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Stage Stable Training Stage] start from step 1 +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: +12/10/2024 06:50:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Start training] datetime: 2024-12-10 06:50:50.697622 | mbs: 8 | grad_accum: 16 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/10/2024 06:50:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/10/2024 06:50:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 1850.43MiB. Peak allocated 5408.00MiB. Peak reserved: 31172.00MiB +12/10/2024 06:51:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 2046.72MiB. Peak allocated 11944.10MiB. Peak reserved: 17106.00MiB +12/10/2024 06:51:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 3.86K | tokens_per_sec: 1.09M | tokens_per_sec_per_gpu: 17K | global_batch_size: 2.05K | lm_loss: 12.2 | lr: 0.00015 | model_tflops_per_gpu: 140 | hardware_tflops_per_gpu: 140 | grad_norm: 0.326 | cuda_memory_allocated: 2.3G | cuda_max_memory_reserved: 18G | hd_total_memory_tb: 312G | hd_used_memory_tb: 67.5G | hd_free_memory_tb: 245G +12/10/2024 06:51:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 2194.80MiB. Peak allocated 3416.07MiB. Peak reserved: 17150.00MiB +12/10/2024 06:51:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 2194.81MiB. Peak allocated 12092.19MiB. Peak reserved: 17158.00MiB +12/10/2024 06:51:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 1.83K | tokens_per_sec: 2.3M | tokens_per_sec_per_gpu: 35.9K | global_batch_size: 2.05K | lm_loss: 12.2 | lr: 0.0003 | model_tflops_per_gpu: 296 | hardware_tflops_per_gpu: 296 | grad_norm: 0.326 | cuda_memory_allocated: 2.3G | cuda_max_memory_reserved: 18G | hd_total_memory_tb: 312G | hd_used_memory_tb: 67.5G | hd_free_memory_tb: 245G +12/10/2024 06:51:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 2194.80MiB. Peak allocated 3416.08MiB. Peak reserved: 17158.00MiB +12/10/2024 06:51:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 2194.81MiB. Peak allocated 12092.19MiB. Peak reserved: 17158.00MiB +num_params +{'total': 1241784320, 'local': 310446080} +num_params +num_paramsnum_params{'total': 1241784320, 'local': 310446080} +num_params + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} + +{'total': 1241784320, 'local': 310446080} +num_params +num_params +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080}num_params + +num_params +num_params{'total': 1241784320, 'local': 310446080} + +num_params +{'total': 1241784320, 'local': 310446080} +num_paramsnum_params + +{'total': 1241784320, 'local': 310446080} +num_params{'total': 1241784320, 'local': 310446080} + +num_paramsnum_paramsnum_params + + +num_params{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080}num_params +{'total': 1241784320, 'local': 310446080}num_params + + + + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080} +num_params +{'total': 1241784320, 'local': 310446080} +num_params +num_params{'total': 1241784320, 'local': 310446080} + +{'total': 1241784320, 'local': 310446080} +num_paramsnum_params +num_params +num_params{'total': 1241784320, 'local': 310446080}num_params +{'total': 1241784320, 'local': 310446080} +num_params + + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080}num_params + +num_params{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080} + + +num_paramsnum_params{'total': 1241784320, 'local': 310446080}num_params + + + +{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080} + +{'total': 1241784320, 'local': 310446080} +num_params + +{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080} + +num_paramsnum_params + +{'total': 1241784320, 'local': 310446080} +num_params{'total': 1241784320, 'local': 310446080} + +num_params{'total': 1241784320, 'local': 310446080} + +num_paramsnum_params + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +num_paramsnum_params{'total': 1241784320, 'local': 310446080} + + +num_params{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080} +num_params + + +num_params +{'total': 1241784320, 'local': 310446080} + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +num_params +{'total': 1241784320, 'local': 310446080} +num_params{'total': 1241784320, 'local': 310446080} + +num_paramsnum_params{'total': 1241784320, 'local': 310446080} + + +num_params +num_params +num_params{'total': 1241784320, 'local': 310446080} + +{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080} + +num_params +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +num_params +{'total': 1241784320, 'local': 310446080} +12/10/2024 06:51:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: iteration: 3 / 100 | consumed_tokens: 12.6M | elapsed_time_per_iteration_ms: 1.82K | tokens_per_sec: 2.3M | tokens_per_sec_per_gpu: 36K | global_batch_size: 2.05K | lm_loss: 12.2 | lr: 0.000296 | model_tflops_per_gpu: 297 | hardware_tflops_per_gpu: 297 | grad_norm: 0.321 | cuda_memory_allocated: 2.3G | cuda_max_memory_reserved: 18G | hd_total_memory_tb: 312G | hd_used_memory_tb: 67.5G | hd_free_memory_tb: 245G +num_params +{'total': 1241784320, 'local': 310446080} +num_params +{'total': 1241784320, 'local': 310446080} +num_paramsnum_params + +num_params{'total': 1241784320, 'local': 310446080}num_params{'total': 1241784320, 'local': 310446080}num_params + +num_params + +num_params + +12/10/2024 06:51:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/10/2024 06:51:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: | -------- | --------------------------------------------------------------- | ----- | ------- | --- | ----------- | ---- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | --------------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080}{'total': 1241784320, 'local': 310446080} + + +{'total': 1241784320, 'local': 310446080} +{'total': 1241784320, 'local': 310446080} +12/10/2024 06:51:06 [INFO|DP=8|PP=0|TP=1|ip-26-0-164-45]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=12|PP=0|TP=1|ip-26-0-165-24]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=6|PP=0|TP=0|ip-26-0-164-236]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=12|PP=0|TP=2|ip-26-0-165-24]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=8|PP=0|TP=0|ip-26-0-164-45]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=9|PP=0|TP=0|ip-26-0-164-45]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=13|PP=0|TP=0|ip-26-0-165-24]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=10|PP=0|TP=0|ip-26-0-164-75]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=9|PP=0|TP=1|ip-26-0-164-45]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=2|PP=0|TP=2|ip-26-0-164-187]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=8|PP=0|TP=3|ip-26-0-164-45]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=9|PP=0|TP=3|ip-26-0-164-45]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=9|PP=0|TP=2|ip-26-0-164-45]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=8|PP=0|TP=2|ip-26-0-164-45]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=13|PP=0|TP=1|ip-26-0-165-24]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=13|PP=0|TP=3|ip-26-0-165-24]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=13|PP=0|TP=2|ip-26-0-165-24]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=5|PP=0|TP=0|ip-26-0-164-207]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=12|PP=0|TP=3|ip-26-0-165-24]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=12|PP=0|TP=0|ip-26-0-165-24]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=3|PP=0|TP=1|ip-26-0-164-187]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=2|PP=0|TP=3|ip-26-0-164-187]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=15|PP=0|TP=0|ip-26-0-165-38]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=3|PP=0|TP=3|ip-26-0-164-187]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=4|PP=0|TP=1|ip-26-0-164-207]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=7|PP=0|TP=0|ip-26-0-164-236]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=4|PP=0|TP=2|ip-26-0-164-207]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=4|PP=0|TP=0|ip-26-0-164-207]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=5|PP=0|TP=2|ip-26-0-164-207]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=5|PP=0|TP=1|ip-26-0-164-207]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=5|PP=0|TP=3|ip-26-0-164-207]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=4|PP=0|TP=3|ip-26-0-164-207]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=10|PP=0|TP=1|ip-26-0-164-75]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=11|PP=0|TP=2|ip-26-0-164-75]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=10|PP=0|TP=2|ip-26-0-164-75]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=11|PP=0|TP=0|ip-26-0-164-75]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=14|PP=0|TP=0|ip-26-0-165-38]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=2|PP=0|TP=0|ip-26-0-164-187]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=3|PP=0|TP=2|ip-26-0-164-187]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=3|PP=0|TP=0|ip-26-0-164-187]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=14|PP=0|TP=2|ip-26-0-165-38]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=15|PP=0|TP=1|ip-26-0-165-38]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=14|PP=0|TP=3|ip-26-0-165-38]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=2|PP=0|TP=1|ip-26-0-164-187]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=15|PP=0|TP=3|ip-26-0-165-38]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=6|PP=0|TP=1|ip-26-0-164-236]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=6|PP=0|TP=3|ip-26-0-164-236]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=6|PP=0|TP=2|ip-26-0-164-236]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=7|PP=0|TP=1|ip-26-0-164-236]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=7|PP=0|TP=3|ip-26-0-164-236]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=7|PP=0|TP=2|ip-26-0-164-236]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=14|PP=0|TP=1|ip-26-0-165-38]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=15|PP=0|TP=2|ip-26-0-165-38]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=11|PP=0|TP=3|ip-26-0-164-75]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=11|PP=0|TP=1|ip-26-0-164-75]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=10|PP=0|TP=3|ip-26-0-164-75]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: | 13265818 | 1.34G_dp16_tp4_pp1_acc16_mbs8_seq2048_zero1_tpmodeRED_vocab131k | 8 | 2048 | 8 | 16 | 2048 | 296.82 | 296.82 | 35958.57 | 218.59 | 173.03 | 168.62 | 458.92 | 264.54 | 263.50 | 3.34 | 16.76 | 16 | 1 | 4 | | TensorParallelLinearMode.REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 8 | 2048 | 131072 | True | torch.bfloat16 | 1 | 25 | True | 1.24G | 310M | +12/10/2024 06:51:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-225]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-225]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-225]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=1|PP=0|TP=1|ip-26-0-160-225]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=1|PP=0|TP=2|ip-26-0-160-225]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=1|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/10/2024 06:51:06 [INFO|DP=1|PP=0|TP=3|ip-26-0-160-225]: Throughput logging complete +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +slurmstepd: error: *** STEP 13265818.0 ON ip-26-0-160-225 CANCELLED AT 2024-12-10T06:51:07 *** +[2024-12-10 06:51:07,540] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26306 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26307 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26308 closing signal SIGTERM +slurmstepd: error: *** JOB 13265818 ON ip-26-0-160-225 CANCELLED AT 2024-12-10T06:51:07 *** +[2024-12-10 06:51:07,543] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177398 closing signal SIGTERM +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177399 closing signal SIGTERM +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177400 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177401 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77796 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77797 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77798 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177402 closing signal SIGTERM +[2024-12-10 06:51:07,540] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85399 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85400 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77799 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26309 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26310 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26311 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26312 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44772 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44773 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44774 closing signal SIGTERM +[2024-12-10 06:51:07,546] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177403 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44775 closing signal SIGTERM +[2024-12-10 06:51:07,541] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44776 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44777 closing signal SIGTERM +[2024-12-10 06:51:07,547] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177404 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77800 closing signal SIGTERM +[2024-12-10 06:51:07,545] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26313 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 06:51:07,548] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177405 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 06:51:07,543] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44778 closing signal SIGTERM +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77801 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 64858 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85401 closing signal SIGTERM +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 821284 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44779 closing signal SIGTERM +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77802 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 64859 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85402 closing signal SIGTERM +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 821285 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77803 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 64860 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85403 closing signal SIGTERM +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 821286 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 64861 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85404 closing signal SIGTERM +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 821287 closing signal SIGTERM +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 64862 closing signal SIGTERM +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 64863 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85405 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 821288 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85406 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 821289 closing signal SIGTERM +[2024-12-10 06:51:07,545] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 821290 closing signal SIGTERM +[2024-12-10 06:51:07,545] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 821291 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3924496 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3924497 closing signal SIGTERM +[2024-12-10 06:51:07,542] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3924498 closing signal SIGTERM +[2024-12-10 06:51:07,543] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3924499 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3924500 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3924501 closing signal SIGTERM +[2024-12-10 06:51:07,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3924502 closing signal SIGTERM +[2024-12-10 06:51:07,545] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3924503 closing signal SIGTERM +[2024-12-10 06:51:07,551] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 64864 closing signal SIGTERM +[2024-12-10 06:51:07,552] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 64865 closing signal SIGTERM diff --git a/logs/13266568-bench_1.14G_dp2_tp8_pp1_acc8_mbs8_seq8192_zero1_tpmodeALL_vocab32k.out b/logs/13266568-bench_1.14G_dp2_tp8_pp1_acc8_mbs8_seq8192_zero1_tpmodeALL_vocab32k.out new file mode 100644 index 0000000000000000000000000000000000000000..e68875bc952f02cee89b761ceaec15aece88583e --- /dev/null +++ b/logs/13266568-bench_1.14G_dp2_tp8_pp1_acc8_mbs8_seq8192_zero1_tpmodeALL_vocab32k.out @@ -0,0 +1,640 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-132,ip-26-0-170-143 ++ export 'NODELIST=ip-26-0-169-132 +ip-26-0-170-143' ++ NODELIST='ip-26-0-169-132 +ip-26-0-170-143' +++ scontrol show hostnames ip-26-0-169-132,ip-26-0-170-143 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-132 ++ MASTER_NODE=ip-26-0-169-132 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-169-132' +Master node: ip-26-0-169-132 ++ echo 'All nodes: ip-26-0-169-132 +ip-26-0-170-143' +All nodes: ip-26-0-169-132 +ip-26-0-170-143 ++ echo 'World size: 16' +World size: 16 ++ srun torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13266568 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-132:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp2_tp8_pp1_acc8_mbs8_seq8192_zero1_tpmodeALL_vocab32k.yaml +[2024-12-10 08:15:28,927] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 08:15:28,927] torch.distributed.run: [WARNING] +[2024-12-10 08:15:28,927] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 08:15:28,927] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 08:15:28,927] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 08:15:29,232] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 08:15:29,232] torch.distributed.run: [WARNING] +[2024-12-10 08:15:29,232] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 08:15:29,232] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 08:15:29,232] torch.distributed.run: [WARNING] ***************************************** +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Config: +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Config(general=GeneralArgs(project='debug', +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: run='1.14G_dp2_tp8_pp1_acc8_mbs8_seq8192_zero1_tpmodeALL_vocab32k', +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: seed=42, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: step=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: consumed_train_samples=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: ignore_sanity_checks=True), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: parallelism=ParallelismArgs(dp=2, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pp=1, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp=8, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pp_engine=, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp_mode=, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp_linear_async_communication=False, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: recompute_layer=False, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp_recompute_allgather=True, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: expert_parallel_size=1), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: eos_token_id=0, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_act='silu', +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_size=2048, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: initializer_range=0.02, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: intermediate_size=8192, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: is_llama_config=True, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: max_position_embeddings=8192, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_attention_heads=32, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_hidden_layers=16, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_key_value_heads=8, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pad_token_id=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pretraining_tp=1, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rms_norm_eps=1e-05, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_scaling=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_theta=10000.0, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_interleaved=False, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tie_word_embeddings=True, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: use_cache=True, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: vocab_size=32768), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: init_method=RandomInit(std=0.02), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: dtype=torch.bfloat16, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: make_vocab_size_divisible_by=1, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: ddp_bucket_cap_mb=25), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokenizer_revision=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokenizer_max_length=None), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: checkpoint_interval=10000, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: save_initial_state=False, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: save_final_state=False, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: resume_checkpoint_path=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: checkpoints_path_is_shared_file_system=False), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: logging=LoggingArgs(log_level='info', +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: log_level_replica='info', +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: iteration_step_info_interval=1), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokens=TokensArgs(sequence_length=8192, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: train_steps=100, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: micro_batch_size=8, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: batch_accumulation_per_replica=8, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: val_check_interval=100, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: limit_val_batches=0, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: limit_test_batches=0), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: adam_beta1=0.9, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: adam_beta2=0.95, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: torch_adam_is_fused=True, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: name='adamW'), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: zero_stage=1, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: weight_decay=0.01, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: clip_grad=1.0, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: accumulate_grad_in_fp32=True, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_warmup_steps=2, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_warmup_style='linear', +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_decay_style='cosine', +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_decay_steps=13, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_decay_starting_step=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: min_decay_lr=1e-05)), +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: start_training_step=1, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: data=DataArgs(dataset=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: seed=42, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_loading_workers=1))], +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: profiler=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lighteval=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: s3_upload=None) +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Model Config: +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: LlamaConfig(bos_token_id=0, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: eos_token_id=0, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_act='silu', +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_size=2048, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: initializer_range=0.02, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: intermediate_size=8192, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: is_llama_config=True, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: max_position_embeddings=8192, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_attention_heads=32, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_hidden_layers=16, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_key_value_heads=8, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pad_token_id=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pretraining_tp=1, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rms_norm_eps=1e-05, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_scaling=None, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_theta=10000.0, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_interleaved=False, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tie_word_embeddings=True, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: use_cache=True, +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: vocab_size=32768) +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Building model.. +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Initialize RoPE Theta = 10000.0 +12/10/2024 08:16:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Setting PP block ranks... +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Total number of parameters: 1.04G (1985.03MiB) +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Local number of parameters: 130M (248.13MiB) +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=2|ip-26-0-169-132]: Local number of parameters: 130M (248.13MiB) +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=4|ip-26-0-169-132]: Local number of parameters: 130M (248.13MiB) +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=6|ip-26-0-169-132]: Local number of parameters: 130M (248.13MiB) +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=5|ip-26-0-169-132]: Local number of parameters: 130M (248.13MiB) +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=7|ip-26-0-169-132]: Local number of parameters: 130M (248.13MiB) +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=3|ip-26-0-169-132]: Local number of parameters: 130M (248.13MiB) +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-132]: Local number of parameters: 130M (248.13MiB) +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [After model building] Memory usage: 249.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35218.00MiB +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=2|ip-26-0-169-132]: [After model building] Memory usage: 249.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35218.00MiB +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=4|ip-26-0-169-132]: [After model building] Memory usage: 249.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35218.00MiB +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=6|ip-26-0-169-132]: [After model building] Memory usage: 249.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35218.00MiB +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=7|ip-26-0-169-132]: [After model building] Memory usage: 249.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35218.00MiB +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=5|ip-26-0-169-132]: [After model building] Memory usage: 249.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35218.00MiB +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=3|ip-26-0-169-132]: [After model building] Memory usage: 249.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35218.00MiB +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-132]: [After model building] Memory usage: 249.15MiB. Peak allocated: 5504.00MiB Peak reserved: 35218.00MiB +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: No checkpoint path provided. +12/10/2024 08:16:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Parametrizing model parameters using StandardParametrizator +12/10/2024 08:16:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [Optimizer Building] Using LearningRateForSP as learning rate +12/10/2024 08:16:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] Size of optimizer params per rank: +12/10/2024 08:16:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] DP Rank 0 has 65M out of 130M (50.00%) params' optimizer states +12/10/2024 08:16:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [ZeRO sharding] DP Rank 1 has 65M out of 130M (50.00%) params' optimizer states +12/10/2024 08:16:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/10/2024 08:16:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Using dummy data generator +12/10/2024 08:16:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [Training Plan] There are 1 training stages +12/10/2024 08:16:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [Stage Stable Training Stage] start from step 1 +12/10/2024 08:16:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: +12/10/2024 08:16:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: [Start training] datetime: 2024-12-10 08:16:15.510311 | mbs: 8 | grad_accum: 8 | global_batch_size: 128 | sequence_length: 8192 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/10/2024 08:16:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/10/2024 08:16:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Memory usage: 993.54MiB. Peak allocated 5504.00MiB. Peak reserved: 35218.00MiB +12/10/2024 08:16:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Memory usage: 1074.67MiB. Peak allocated 30030.32MiB. Peak reserved: 42420.00MiB +12/10/2024 08:16:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 4.61K | tokens_per_sec: 228K | tokens_per_sec_per_gpu: 14.2K | global_batch_size: 128 | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 135 | hardware_tflops_per_gpu: 135 | grad_norm: 0.479 | cuda_memory_allocated: 1.66G | cuda_max_memory_reserved: 44.5G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.8G | hd_free_memory_tb: 241G +12/10/2024 08:16:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Memory usage: 1582.48MiB. Peak allocated 2202.80MiB. Peak reserved: 42432.00MiB +12/10/2024 08:16:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Memory usage: 1582.48MiB. Peak allocated 30538.13MiB. Peak reserved: 42452.00MiB +12/10/2024 08:16:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 2.3K | tokens_per_sec: 457K | tokens_per_sec_per_gpu: 28.5K | global_batch_size: 128 | lm_loss: 10.8 | lr: 0.0003 | model_tflops_per_gpu: 270 | hardware_tflops_per_gpu: 270 | grad_norm: 0.481 | cuda_memory_allocated: 1.66G | cuda_max_memory_reserved: 44.5G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.8G | hd_free_memory_tb: 241G +12/10/2024 08:16:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Memory usage: 1582.48MiB. Peak allocated 2202.80MiB. Peak reserved: 42452.00MiB +12/10/2024 08:16:32 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Memory usage: 1582.48MiB. Peak allocated 30538.13MiB. Peak reserved: 42452.00MiB +num_paramsnum_paramsnum_params + +num_params{'total': 1040728064, 'local': 130091008}{'total': 1040728064, 'local': 130091008} + + + +{'total': 1040728064, 'local': 130091008} +{'total': 1040728064, 'local': 130091008} +num_params +num_params{'total': 1040728064, 'local': 130091008} + +{'total': 1040728064, 'local': 130091008} +num_params +{'total': 1040728064, 'local': 130091008} +num_params +{'total': 1040728064, 'local': 130091008}num_params + +{'total': 1040728064, 'local': 130091008} +num_paramsnum_paramsnum_params + +{'total': 1040728064, 'local': 130091008} +{'total': 1040728064, 'local': 130091008} +num_params{'total': 1040728064, 'local': 130091008} +num_params + + +{'total': 1040728064, 'local': 130091008} +{'total': 1040728064, 'local': 130091008} +num_params +{'total': 1040728064, 'local': 130091008} +12/10/2024 08:16:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 2.3K | tokens_per_sec: 457K | tokens_per_sec_per_gpu: 28.5K | global_batch_size: 128 | lm_loss: 10.8 | lr: 0.000296 | model_tflops_per_gpu: 270 | hardware_tflops_per_gpu: 270 | grad_norm: 0.464 | cuda_memory_allocated: 1.66G | cuda_max_memory_reserved: 44.5G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.8G | hd_free_memory_tb: 241G +num_params +{'total': 1040728064, 'local': 130091008} +12/10/2024 08:16:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/10/2024 08:16:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: | -------- | ------------------------------------------------------------ | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | ----------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/10/2024 08:16:33 [INFO|DP=1|PP=0|TP=4|ip-26-0-170-143]: Throughput logging complete +12/10/2024 08:16:33 [INFO|DP=1|PP=0|TP=5|ip-26-0-170-143]: Throughput logging complete +12/10/2024 08:16:33 [INFO|DP=1|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +12/10/2024 08:16:33 [INFO|DP=1|PP=0|TP=6|ip-26-0-170-143]: Throughput logging complete +12/10/2024 08:16:33 [INFO|DP=1|PP=0|TP=1|ip-26-0-170-143]: Throughput logging complete +12/10/2024 08:16:33 [INFO|DP=1|PP=0|TP=3|ip-26-0-170-143]: Throughput logging complete +12/10/2024 08:16:33 [INFO|DP=1|PP=0|TP=2|ip-26-0-170-143]: Throughput logging complete +12/10/2024 08:16:33 [INFO|DP=1|PP=0|TP=7|ip-26-0-170-143]: Throughput logging complete +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +slurmstepd: error: *** JOB 13266568 ON ip-26-0-169-132 CANCELLED AT 2024-12-10T08:16:33 *** +[2024-12-10 08:16:33,047] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-10 08:16:33,047] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1758126 closing signal SIGTERM +[2024-12-10 08:16:33,047] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1758127 closing signal SIGTERM +[2024-12-10 08:16:33,047] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1758128 closing signal SIGTERM +[2024-12-10 08:16:33,047] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1758129 closing signal SIGTERM +[2024-12-10 08:16:33,047] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1758130 closing signal SIGTERM diff --git a/logs/13290958-bench_1.14G_dp24_tp1_pp1_acc1_mbs8_seq2048_zero1_tpmodeRED_vocab32k_prof.out b/logs/13290958-bench_1.14G_dp24_tp1_pp1_acc1_mbs8_seq2048_zero1_tpmodeRED_vocab32k_prof.out new file mode 100644 index 0000000000000000000000000000000000000000..2601c0bb1101f3f09458bf773ce2fcfc7d76920f --- /dev/null +++ b/logs/13290958-bench_1.14G_dp24_tp1_pp1_acc1_mbs8_seq2048_zero1_tpmodeRED_vocab32k_prof.out @@ -0,0 +1,829 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-168-238,ip-26-0-169-[132,207]' ++ export 'NODELIST=ip-26-0-168-238 +ip-26-0-169-132 +ip-26-0-169-207' ++ NODELIST='ip-26-0-168-238 +ip-26-0-169-132 +ip-26-0-169-207' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-168-238,ip-26-0-169-[132,207]' ++ export MASTER_NODE=ip-26-0-168-238 ++ MASTER_NODE=ip-26-0-168-238 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=3 ++ NNODES=3 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=24 ++ WORLD_SIZE=24 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-168-238' +Master node: ip-26-0-168-238 ++ echo 'All nodes: ip-26-0-168-238 +ip-26-0-169-132 +ip-26-0-169-207' +All nodes: ip-26-0-168-238 +ip-26-0-169-132 +ip-26-0-169-207 ++ echo 'World size: 24' +World size: 24 ++ srun torchrun --nnodes=3 --nproc_per_node=8 --rdzv_id=13290958 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-168-238:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp24_tp1_pp1_acc1_mbs8_seq2048_zero1_tpmodeRED_vocab32k_prof.yaml +[2024-12-10 16:18:19,930] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 16:18:19,930] torch.distributed.run: [WARNING] +[2024-12-10 16:18:19,930] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 16:18:19,930] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 16:18:19,930] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 16:18:19,974] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 16:18:19,974] torch.distributed.run: [WARNING] +[2024-12-10 16:18:19,974] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 16:18:19,974] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 16:18:19,974] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 16:18:20,036] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-10 16:18:20,037] torch.distributed.run: [WARNING] +[2024-12-10 16:18:20,037] torch.distributed.run: [WARNING] ***************************************** +[2024-12-10 16:18:20,037] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-10 16:18:20,037] torch.distributed.run: [WARNING] ***************************************** +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Config: +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Config(general=GeneralArgs(project='debug', +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: run='1.14G_dp24_tp1_pp1_acc1_mbs8_seq2048_zero1_tpmodeRED_vocab32k_prof', +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: seed=42, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: step=None, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: consumed_train_samples=None, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: ignore_sanity_checks=True), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: parallelism=ParallelismArgs(dp=24, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: pp=1, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: tp=1, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: pp_engine=, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: tp_mode=, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: tp_linear_async_communication=True, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: recompute_layer=False, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: tp_recompute_allgather=True, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: expert_parallel_size=1), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: eos_token_id=0, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: hidden_act='silu', +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: hidden_size=2048, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: initializer_range=0.02, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: intermediate_size=8192, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: is_llama_config=True, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: max_position_embeddings=2048, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: num_attention_heads=32, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: num_hidden_layers=16, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: num_key_value_heads=8, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: pad_token_id=None, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: pretraining_tp=1, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: rms_norm_eps=1e-05, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: rope_scaling=None, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: rope_theta=10000.0, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: rope_interleaved=False, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: tie_word_embeddings=True, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: use_cache=True, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: vocab_size=32768), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: init_method=RandomInit(std=0.02), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: dtype=torch.bfloat16, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: make_vocab_size_divisible_by=1, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: ddp_bucket_cap_mb=25), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: tokenizer_revision=None, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: tokenizer_max_length=None), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: checkpoint_interval=10000, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: save_initial_state=False, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: save_final_state=False, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: resume_checkpoint_path=None, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: checkpoints_path_is_shared_file_system=False), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: logging=LoggingArgs(log_level='info', +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: log_level_replica='info', +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: iteration_step_info_interval=1), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: tokens=TokensArgs(sequence_length=2048, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: train_steps=10, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: micro_batch_size=8, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: batch_accumulation_per_replica=1, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: val_check_interval=100, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: limit_val_batches=0, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: limit_test_batches=0), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: adam_beta1=0.9, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: adam_beta2=0.95, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: torch_adam_is_fused=True, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: name='adamW'), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: zero_stage=1, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: weight_decay=0.01, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: clip_grad=1.0, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: accumulate_grad_in_fp32=True, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: lr_warmup_steps=2, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: lr_warmup_style='linear', +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: lr_decay_style='cosine', +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: lr_decay_steps=13, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: lr_decay_starting_step=None, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: min_decay_lr=1e-05)), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: start_training_step=1, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: data=DataArgs(dataset=None, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: seed=42, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: num_loading_workers=1))], +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: profiler=ProfilerArgs(profiler_export_path=PosixPath('tb_logs')), +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: lighteval=None, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: s3_upload=None) +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Model Config: +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: LlamaConfig(bos_token_id=0, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: eos_token_id=0, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: hidden_act='silu', +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: hidden_size=2048, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: initializer_range=0.02, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: intermediate_size=8192, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: is_llama_config=True, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: max_position_embeddings=2048, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: num_attention_heads=32, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: num_hidden_layers=16, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: num_key_value_heads=8, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: pad_token_id=None, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: pretraining_tp=1, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: rms_norm_eps=1e-05, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: rope_scaling=None, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: rope_theta=10000.0, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: rope_interleaved=False, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: tie_word_embeddings=True, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: use_cache=True, +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: vocab_size=32768) +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Building model.. +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Initialize RoPE Theta = 10000.0 +12/10/2024 16:18:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Setting PP block ranks... +12/10/2024 16:18:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Total number of parameters: 1.04G (1984.13MiB) +12/10/2024 16:18:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Local number of parameters: 1.04G (1984.13MiB) +12/10/2024 16:18:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [After model building] Memory usage: 1984.15MiB. Peak allocated: 2112.15MiB Peak reserved: 2126.00MiB +12/10/2024 16:18:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: No checkpoint path provided. +12/10/2024 16:18:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Parametrizing model parameters using StandardParametrizator +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [Optimizer Building] Using LearningRateForSP as learning rate +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] Size of optimizer params per rank: +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 0 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 1 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 2 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 3 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 4 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 5 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 6 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 7 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 8 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 9 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 10 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 11 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 12 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 13 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 14 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 15 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 16 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 17 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 18 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 19 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 20 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 21 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 22 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [ZeRO sharding] DP Rank 23 has 43.3M out of 1.04G (4.17%) params' optimizer states +12/10/2024 16:19:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [Training Plan] Stage Stable Training Stage has 9 remaining training steps and has consumed 0 samples +12/10/2024 16:19:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Using dummy data generator +12/10/2024 16:19:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [Training Plan] There are 1 training stages +12/10/2024 16:19:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [Stage Stable Training Stage] start from step 1 +12/10/2024 16:19:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: +12/10/2024 16:19:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: [Start training] datetime: 2024-12-10 16:19:04.499166 | mbs: 8 | grad_accum: 1 | global_batch_size: 192 | sequence_length: 2048 | train_steps: 10 | start_iteration_step: 0 | consumed_train_samples: 0 +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +12/10/2024 16:19:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 9 remaining train steps +12/10/2024 16:19:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Memory usage: 6118.41MiB. Peak allocated 6118.41MiB. Peak reserved: 6262.00MiB +12/10/2024 16:19:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Memory usage: 6186.69MiB. Peak allocated 33453.48MiB. Peak reserved: 34642.00MiB +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +12/10/2024 16:19:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: iteration: 1 / 10 | consumed_tokens: 393K | elapsed_time_per_iteration_ms: 2.36K | tokens_per_sec: 167K | tokens_per_sec_per_gpu: 6.94K | global_batch_size: 192 | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 48.9 | hardware_tflops_per_gpu: 48.9 | grad_norm: 1.89 | cuda_memory_allocated: 6.83G | cuda_max_memory_reserved: 36.4G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.2G | hd_free_memory_tb: 242G +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +12/10/2024 16:19:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Memory usage: 6517.47MiB. Peak allocated 8584.28MiB. Peak reserved: 34706.00MiB +12/10/2024 16:19:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Memory usage: 6517.47MiB. Peak allocated 33819.58MiB. Peak reserved: 34706.00MiB +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +12/10/2024 16:19:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: iteration: 2 / 10 | consumed_tokens: 786K | elapsed_time_per_iteration_ms: 330 | tokens_per_sec: 1.19M | tokens_per_sec_per_gpu: 49.6K | global_batch_size: 192 | lm_loss: 10.8 | lr: 0.0003 | model_tflops_per_gpu: 349 | hardware_tflops_per_gpu: 349 | grad_norm: 1.89 | cuda_memory_allocated: 6.83G | cuda_max_memory_reserved: 36.4G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.2G | hd_free_memory_tb: 242G +[W kineto_shim.cpp:372] Profiler is not initialized: skipping step() invocation +12/10/2024 16:19:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Memory usage: 6517.47MiB. Peak allocated 8584.28MiB. Peak reserved: 34706.00MiB +12/10/2024 16:19:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Memory usage: 6517.47MiB. Peak allocated 33819.58MiB. Peak reserved: 34706.00MiB +12/10/2024 16:19:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: iteration: 3 / 10 | consumed_tokens: 1.18M | elapsed_time_per_iteration_ms: 330 | tokens_per_sec: 1.19M | tokens_per_sec_per_gpu: 49.6K | global_batch_size: 192 | lm_loss: 10.8 | lr: 0.000296 | model_tflops_per_gpu: 350 | hardware_tflops_per_gpu: 350 | grad_norm: 1.85 | cuda_memory_allocated: 6.83G | cuda_max_memory_reserved: 36.4G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.2G | hd_free_memory_tb: 242G +12/10/2024 16:19:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/10/2024 16:19:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: | -------- | ------------------------------------------------------------------ | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | --------------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/10/2024 16:19:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: | 13290958 | 1.14G_dp24_tp1_pp1_acc1_mbs8_seq2048_zero1_tpmodeRED_vocab32k_prof | 3 | 2048 | 8 | 1 | 192 | 349.72 | 349.72 | 49630.25 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 8.38 | 33.89 | 24 | 1 | 1 | | TensorParallelLinearMode.REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 8 | 2048 | 32768 | True | torch.bfloat16 | 1 | 25 | True | 1.04G | 1.04G | +12/10/2024 16:19:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=6|PP=0|TP=0|ip-26-0-168-238]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=5|PP=0|TP=0|ip-26-0-168-238]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=1|PP=0|TP=0|ip-26-0-168-238]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=3|PP=0|TP=0|ip-26-0-168-238]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=2|PP=0|TP=0|ip-26-0-168-238]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=4|PP=0|TP=0|ip-26-0-168-238]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=18|PP=0|TP=0|ip-26-0-169-207]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=23|PP=0|TP=0|ip-26-0-169-207]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=16|PP=0|TP=0|ip-26-0-169-207]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=7|PP=0|TP=0|ip-26-0-168-238]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=17|PP=0|TP=0|ip-26-0-169-207]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=21|PP=0|TP=0|ip-26-0-169-207]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=19|PP=0|TP=0|ip-26-0-169-207]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=20|PP=0|TP=0|ip-26-0-169-207]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=22|PP=0|TP=0|ip-26-0-169-207]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=8|PP=0|TP=0|ip-26-0-169-132]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=13|PP=0|TP=0|ip-26-0-169-132]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=14|PP=0|TP=0|ip-26-0-169-132]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=15|PP=0|TP=0|ip-26-0-169-132]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=12|PP=0|TP=0|ip-26-0-169-132]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=10|PP=0|TP=0|ip-26-0-169-132]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=11|PP=0|TP=0|ip-26-0-169-132]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=9|PP=0|TP=0|ip-26-0-169-132]: Throughput logging complete +12/10/2024 16:19:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Memory usage: 6517.47MiB. Peak allocated 8584.28MiB. Peak reserved: 34706.00MiB +12/10/2024 16:19:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Memory usage: 6517.47MiB. Peak allocated 33819.58MiB. Peak reserved: 34706.00MiB +STAGE:2024-12-10 16:19:17 21730:21730 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 21731:21731 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 21735:21735 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 21734:21734 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 21733:21733 ActivityProfilerController.cpp:312] Completed Stage: Warm Up + + + + +STAGE:2024-12-10 16:19:17 80774:80774 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 80778:80778 ActivityProfilerController.cpp:312] Completed Stage: Warm Up + +12/10/2024 16:19:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: iteration: 4 / 10 | consumed_tokens: 1.57M | elapsed_time_per_iteration_ms: 348 | tokens_per_sec: 1.13M | tokens_per_sec_per_gpu: 47.1K | global_batch_size: 192 | lm_loss: 10.8 | lr: 0.000283 | model_tflops_per_gpu: 332 | hardware_tflops_per_gpu: 332 | grad_norm: 1.67 | cuda_memory_allocated: 6.83G | cuda_max_memory_reserved: 36.4G | hd_total_memory_tb: 312G | hd_used_memory_tb: 70.2G | hd_free_memory_tb: 242G +STAGE:2024-12-10 16:19:17 21736:21736 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 21732:21732 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 21737:21737 ActivityProfilerController.cpp:312] Completed Stage: Warm Up + + +STAGE:2024-12-10 16:19:17 936268:936268 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 936267:936267 ActivityProfilerController.cpp:312] Completed Stage: Warm Up +STAGE:2024-12-10 16:19:17 936265:936265 ActivityProfilerController.cpp:312] Completed Stage: Warm Up +STAGE:2024-12-10 16:19:17 936266:936266 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 936261:936261 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 936264:936264 ActivityProfilerController.cpp:312] Completed Stage: Warm Up + +STAGE:2024-12-10 16:19:17 80776:80776 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 80779:80779 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 80775:80775 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 80777:80777 ActivityProfilerController.cpp:312] Completed Stage: Warm UpSTAGE:2024-12-10 16:19:17 80773:80773 ActivityProfilerController.cpp:312] Completed Stage: Warm Up + + + + + +STAGE:2024-12-10 16:19:17 936262:936262 ActivityProfilerController.cpp:312] Completed Stage: Warm Up +STAGE:2024-12-10 16:19:17 936263:936263 ActivityProfilerController.cpp:312] Completed Stage: Warm Up + +STAGE:2024-12-10 16:19:17 80772:80772 ActivityProfilerController.cpp:312] Completed Stage: Warm Up +12/10/2024 16:19:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Memory usage: 6517.47MiB. Peak allocated 8584.28MiB. Peak reserved: 34706.00MiB +12/10/2024 16:19:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: Memory usage: 6517.47MiB. Peak allocated 33819.58MiB. Peak reserved: 34706.00MiB +12/10/2024 16:19:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: iteration: 5 / 10 | consumed_tokens: 1.97M | elapsed_time_per_iteration_ms: 362 | tokens_per_sec: 1.09M | tokens_per_sec_per_gpu: 45.3K | global_batch_size: 192 | lm_loss: 10.8 | lr: 0.000264 | model_tflops_per_gpu: 319 | hardware_tflops_per_gpu: 319 | grad_norm: 1.35 +STAGE:2024-12-10 16:19:17 21736:21736 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 21737:21737 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 21732:21732 ActivityProfilerController.cpp:318] Completed Stage: CollectionSTAGE:2024-12-10 16:19:17 21735:21735 ActivityProfilerController.cpp:318] Completed Stage: Collection + +STAGE:2024-12-10 16:19:17 21730:21730 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 80775:80775 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 936266:936266 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 80773:80773 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 936268:936268 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 80779:80779 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 936265:936265 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 936267:936267 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 936264:936264 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 80776:80776 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 80777:80777 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 936263:936263 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 80778:80778 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 936261:936261 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 80774:80774 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 21731:21731 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 80772:80772 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 21733:21733 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 21734:21734 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 936262:936262 ActivityProfilerController.cpp:318] Completed Stage: Collection +STAGE:2024-12-10 16:19:17 21730:21730 ActivityProfilerController.cpp:322] Completed Stage: Post ProcessingSTAGE:2024-12-10 16:19:17 21735:21735 ActivityProfilerController.cpp:322] Completed Stage: Post Processing + +STAGE:2024-12-10 16:19:17 21736:21736 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 21732:21732 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 21737:21737 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 80779:80779 ActivityProfilerController.cpp:322] Completed Stage: Post ProcessingSTAGE:2024-12-10 16:19:17 80775:80775 ActivityProfilerController.cpp:322] Completed Stage: Post Processing + +STAGE:2024-12-10 16:19:17 80773:80773 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 936267:936267 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 936266:936266 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 936265:936265 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 936268:936268 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 936264:936264 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 936263:936263 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 80778:80778 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 80776:80776 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 936261:936261 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 80777:80777 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 80774:80774 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 21731:21731 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 21734:21734 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 80772:80772 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 21733:21733 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +STAGE:2024-12-10 16:19:17 936262:936262 ActivityProfilerController.cpp:322] Completed Stage: Post Processing +12/10/2024 16:19:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: iteration: 6 / 10 | consumed_tokens: 2.36M | elapsed_time_per_iteration_ms: 476 | tokens_per_sec: 826K | tokens_per_sec_per_gpu: 34.4K | global_batch_size: 192 | lm_loss: 10.7 | lr: 0.000237 | model_tflops_per_gpu: 243 | hardware_tflops_per_gpu: 243 | grad_norm: 0.578 +12/10/2024 16:19:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: iteration: 7 / 10 | consumed_tokens: 2.75M | elapsed_time_per_iteration_ms: 334 | tokens_per_sec: 1.18M | tokens_per_sec_per_gpu: 49K | global_batch_size: 192 | lm_loss: 10.7 | lr: 0.000206 | model_tflops_per_gpu: 345 | hardware_tflops_per_gpu: 345 | grad_norm: 0.387 +12/10/2024 16:19:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: iteration: 8 / 10 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 336 | tokens_per_sec: 1.17M | tokens_per_sec_per_gpu: 48.8K | global_batch_size: 192 | lm_loss: 10.6 | lr: 0.000172 | model_tflops_per_gpu: 344 | hardware_tflops_per_gpu: 344 | grad_norm: 0.223 +12/10/2024 16:19:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: iteration: 9 / 10 | consumed_tokens: 3.54M | elapsed_time_per_iteration_ms: 336 | tokens_per_sec: 1.17M | tokens_per_sec_per_gpu: 48.8K | global_batch_size: 192 | lm_loss: 10.5 | lr: 0.000138 | model_tflops_per_gpu: 344 | hardware_tflops_per_gpu: 344 | grad_norm: 0.174 +12/10/2024 16:19:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-238]: iteration: 10 / 10 | consumed_tokens: 3.93M | elapsed_time_per_iteration_ms: 336 | tokens_per_sec: 1.17M | tokens_per_sec_per_gpu: 48.8K | global_batch_size: 192 | lm_loss: 10.5 | lr: 0.000104 | model_tflops_per_gpu: 344 | hardware_tflops_per_gpu: 344 | grad_norm: 0.147 +wandb: - 0.000 MB of 0.000 MB uploaded wandb: +wandb: Run history: +wandb: consumed_tokens ▁▂▃▃▄▅▆▆▇█ +wandb: cuda_max_memory_reserved ▁▁▁▁ +wandb: cuda_memory_allocated ▁▁▁▁ +wandb: elapsed_time_per_iteration_ms █▁▁▁▁▂▁▁▁▁ +wandb: global_batch_size ▁▁▁▁▁▁▁▁▁▁ +wandb: grad_norm ███▇▆▃▂▁▁▁ +wandb: hardware_tflops_per_gpu ▁███▇▆████ +wandb: hd_free_memory_tb █▅▅▁ +wandb: hd_total_memory_tb ▁▁▁▁ +wandb: hd_used_memory_tb ▁▅▅█ +wandb: iteration_step ▁▂▃▃▄▅▆▆▇█ +wandb: lm_loss █████▆▅▃▂▁ +wandb: lr ▃██▇▇▆▅▃▂▁ +wandb: model_tflops_per_gpu ▁███▇▆████ +wandb: tokens_per_sec ▁███▇▆████ +wandb: tokens_per_sec_per_gpu ▁███▇▆████ +wandb: +wandb: Run summary: +wandb: consumed_tokens 3932160 +wandb: cuda_max_memory_reserved 36391878656 +wandb: cuda_memory_allocated 6834059776 +wandb: elapsed_time_per_iteration_ms 335.57439 +wandb: global_batch_size 192 +wandb: grad_norm 0.14707 +wandb: hardware_tflops_per_gpu 344.03317 +wandb: hd_free_memory_tb 241984868352 +wandb: hd_total_memory_tb 312187465728 +wandb: hd_used_memory_tb 70185820160 +wandb: iteration_step 10 +wandb: lm_loss 10.48225 +wandb: lr 0.0001 +wandb: model_tflops_per_gpu 344.03317 +wandb: tokens_per_sec 1171769.99637 +wandb: tokens_per_sec_per_gpu 48823.74985 +wandb: +wandb: You can sync this run to the cloud by running: +wandb: wandb sync /tmp/wandb/offline-run-20241210_161906-pgfab17v +wandb: Find logs at: /tmp/wandb/offline-run-20241210_161906-pgfab17v/logs diff --git a/logs/13385372-bench_stress_test.out b/logs/13385372-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..951a073105660b4f57c854aaabb337e2b30ca5b8 --- /dev/null +++ b/logs/13385372-bench_stress_test.out @@ -0,0 +1,602 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-160-103 ++ export NODELIST=ip-26-0-160-103 ++ NODELIST=ip-26-0-160-103 +++ scontrol show hostnames ip-26-0-160-103 +++ head -n1 ++ export MASTER_NODE=ip-26-0-160-103 ++ MASTER_NODE=ip-26-0-160-103 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-103' +Master node: ip-26-0-160-103 ++ echo 'All nodes: ip-26-0-160-103' +All nodes: ip-26-0-160-103 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13385372 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-103:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-12 16:38:43,201] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-12 16:38:43,201] torch.distributed.run: [WARNING] +[2024-12-12 16:38:43,201] torch.distributed.run: [WARNING] ***************************************** +[2024-12-12 16:38:43,201] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-12 16:38:43,201] torch.distributed.run: [WARNING] ***************************************** +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Config: +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Config(general=GeneralArgs(project='debug', +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: run='1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k', +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: seed=42, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: step=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: consumed_train_samples=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: ignore_sanity_checks=True), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: parallelism=ParallelismArgs(dp=4, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pp=1, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp=2, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pp_engine=, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_mode=, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_linear_async_communication=True, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: recompute_layer=False, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_recompute_allgather=True, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: expert_parallel_size=1), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: eos_token_id=0, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_act='silu', +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_size=2048, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: initializer_range=0.02, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: intermediate_size=8192, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: is_llama_config=True, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: max_position_embeddings=2048, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_attention_heads=32, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_hidden_layers=16, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_key_value_heads=8, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pad_token_id=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pretraining_tp=1, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rms_norm_eps=1e-05, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_scaling=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_theta=10000.0, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_interleaved=False, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tie_word_embeddings=True, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: use_cache=True, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: vocab_size=32768), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: init_method=RandomInit(std=0.02), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: dtype=torch.bfloat16, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: make_vocab_size_divisible_by=1, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: ddp_bucket_cap_mb=25), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer_revision=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer_max_length=None), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoint_interval=10000, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: save_initial_state=False, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: save_final_state=False, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: resume_checkpoint_path=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoints_path_is_shared_file_system=False), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: logging=LoggingArgs(log_level='info', +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: log_level_replica='info', +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration_step_info_interval=1), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokens=TokensArgs(sequence_length=2048, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: train_steps=100, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: micro_batch_size=2, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: batch_accumulation_per_replica=256, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: val_check_interval=100, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: limit_val_batches=0, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: limit_test_batches=0), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: adam_beta1=0.9, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: adam_beta2=0.95, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: torch_adam_is_fused=True, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: name='adamW'), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: zero_stage=1, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: weight_decay=0.01, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: clip_grad=1.0, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: accumulate_grad_in_fp32=True, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_warmup_steps=2, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_warmup_style='linear', +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_style='cosine', +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_steps=13, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_starting_step=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: min_decay_lr=1e-05)), +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: start_training_step=1, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: data=DataArgs(dataset=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: seed=42, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_loading_workers=1))], +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: profiler=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lighteval=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: s3_upload=None) +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Model Config: +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: LlamaConfig(bos_token_id=0, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: eos_token_id=0, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_act='silu', +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_size=2048, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: initializer_range=0.02, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: intermediate_size=8192, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: is_llama_config=True, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: max_position_embeddings=2048, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_attention_heads=32, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_hidden_layers=16, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_key_value_heads=8, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pad_token_id=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pretraining_tp=1, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rms_norm_eps=1e-05, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_scaling=None, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_theta=10000.0, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_interleaved=False, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tie_word_embeddings=True, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: use_cache=True, +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: vocab_size=32768) +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Building model.. +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Initialize RoPE Theta = 10000.0 +12/12/2024 16:39:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Setting PP block ranks... +12/12/2024 16:39:20 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-103]: Local number of parameters: 520M (992.13MiB) +12/12/2024 16:39:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Total number of parameters: 1.04G (1984.26MiB) +12/12/2024 16:39:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Local number of parameters: 520M (992.13MiB) +12/12/2024 16:39:20 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-103]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/12/2024 16:39:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 19970.00MiB +12/12/2024 16:39:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: No checkpoint path provided. +12/12/2024 16:39:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Parametrizing model parameters using StandardParametrizator +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Optimizer Building] Using LearningRateForSP as learning rate +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] Size of optimizer params per rank: +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 0 has 130M out of 520M (25.00%) params' optimizer states +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 1 has 130M out of 520M (25.00%) params' optimizer states +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 2 has 130M out of 520M (25.00%) params' optimizer states +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 3 has 130M out of 520M (25.00%) params' optimizer states +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Using dummy data generator +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Training Plan] There are 1 training stages +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Stage Stable Training Stage] start from step 1 +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: +12/12/2024 16:39:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Start training] datetime: 2024-12-12 16:39:25.897910 | mbs: 2 | grad_accum: 256 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/12/2024 16:39:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/12/2024 16:39:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 3472.48MiB. Peak allocated 5632.00MiB. Peak reserved: 21956.00MiB +12/12/2024 16:39:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 3572.67MiB. Peak allocated 6985.16MiB. Peak reserved: 7434.00MiB +12/12/2024 16:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 20.8K | tokens_per_sec: 202K | tokens_per_sec_per_gpu: 25.2K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 178 | hardware_tflops_per_gpu: 178 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.62G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.4G | hd_free_memory_tb: 243G +12/12/2024 16:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4587.85MiB. Peak allocated 6820.14MiB. Peak reserved: 9178.00MiB +12/12/2024 16:40:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4587.97MiB. Peak allocated 8000.47MiB. Peak reserved: 9178.00MiB +12/12/2024 16:40:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 19.2K | tokens_per_sec: 219K | tokens_per_sec_per_gpu: 27.3K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.0003 | model_tflops_per_gpu: 193 | hardware_tflops_per_gpu: 193 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.62G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.4G | hd_free_memory_tb: 243G +12/12/2024 16:40:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4587.85MiB. Peak allocated 6820.26MiB. Peak reserved: 9178.00MiB +slurmstepd: error: *** STEP 13385372.0 ON ip-26-0-160-103 CANCELLED AT 2024-12-12T16:40:18 *** +[2024-12-12 16:40:18,323] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-12 16:40:18,324] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1851870 closing signal SIGTERM +[2024-12-12 16:40:18,324] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1851871 closing signal SIGTERM +[2024-12-12 16:40:18,324] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1851872 closing signal SIGTERM +[2024-12-12 16:40:18,324] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1851873 closing signal SIGTERM +[2024-12-12 16:40:18,324] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1851874 closing signal SIGTERM +[2024-12-12 16:40:18,326] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1851875 closing signal SIGTERM +[2024-12-12 16:40:18,327] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1851876 closing signal SIGTERM +[2024-12-12 16:40:18,328] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1851877 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1851863 got signal: 15 +srun: error: ip-26-0-160-103: task 0: Exited with exit code 1 diff --git a/logs/13385524-bench_stress_test.out b/logs/13385524-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..35d5627463651befd4250d1ec09f485909dd8acb --- /dev/null +++ b/logs/13385524-bench_stress_test.out @@ -0,0 +1,586 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-160-103 ++ export NODELIST=ip-26-0-160-103 ++ NODELIST=ip-26-0-160-103 +++ scontrol show hostnames ip-26-0-160-103 +++ head -n1 ++ export MASTER_NODE=ip-26-0-160-103 ++ MASTER_NODE=ip-26-0-160-103 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-103' +Master node: ip-26-0-160-103 ++ echo 'All nodes: ip-26-0-160-103' +All nodes: ip-26-0-160-103 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13385524 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-103:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-12 17:28:44,189] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-12 17:28:44,189] torch.distributed.run: [WARNING] +[2024-12-12 17:28:44,189] torch.distributed.run: [WARNING] ***************************************** +[2024-12-12 17:28:44,189] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-12 17:28:44,189] torch.distributed.run: [WARNING] ***************************************** +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Config: +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Config(general=GeneralArgs(project='debug', +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: run='1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k', +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: seed=42, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: step=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: consumed_train_samples=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: ignore_sanity_checks=True), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: parallelism=ParallelismArgs(dp=4, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pp=1, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp=2, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pp_engine=, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_mode=, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_linear_async_communication=True, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: recompute_layer=False, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_recompute_allgather=True, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: expert_parallel_size=1), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: eos_token_id=0, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_act='silu', +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_size=2048, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: initializer_range=0.02, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: intermediate_size=8192, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: is_llama_config=True, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: max_position_embeddings=2048, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_attention_heads=32, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_hidden_layers=16, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_key_value_heads=8, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pad_token_id=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pretraining_tp=1, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rms_norm_eps=1e-05, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_scaling=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_theta=10000.0, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_interleaved=False, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tie_word_embeddings=True, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: use_cache=True, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: vocab_size=32768), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: init_method=RandomInit(std=0.02), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: dtype=torch.bfloat16, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: make_vocab_size_divisible_by=1, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: ddp_bucket_cap_mb=25), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer_revision=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer_max_length=None), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoint_interval=10000, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: save_initial_state=False, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: save_final_state=False, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: resume_checkpoint_path=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoints_path_is_shared_file_system=False), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: logging=LoggingArgs(log_level='info', +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: log_level_replica='info', +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration_step_info_interval=1), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokens=TokensArgs(sequence_length=2048, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: train_steps=100, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: micro_batch_size=2, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: batch_accumulation_per_replica=256, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: val_check_interval=100, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: limit_val_batches=0, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: limit_test_batches=0), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: adam_beta1=0.9, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: adam_beta2=0.95, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: torch_adam_is_fused=True, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: name='adamW'), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: zero_stage=1, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: weight_decay=0.01, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: clip_grad=1.0, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: accumulate_grad_in_fp32=True, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_warmup_steps=2, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_warmup_style='linear', +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_style='cosine', +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_steps=13, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_starting_step=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: min_decay_lr=1e-05)), +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: start_training_step=1, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: data=DataArgs(dataset=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: seed=42, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_loading_workers=1))], +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: profiler=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lighteval=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: s3_upload=None) +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Model Config: +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: LlamaConfig(bos_token_id=0, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: eos_token_id=0, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_act='silu', +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_size=2048, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: initializer_range=0.02, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: intermediate_size=8192, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: is_llama_config=True, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: max_position_embeddings=2048, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_attention_heads=32, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_hidden_layers=16, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_key_value_heads=8, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pad_token_id=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pretraining_tp=1, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rms_norm_eps=1e-05, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_scaling=None, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_theta=10000.0, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_interleaved=False, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tie_word_embeddings=True, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: use_cache=True, +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: vocab_size=32768) +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Building model.. +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Initialize RoPE Theta = 10000.0 +12/12/2024 17:29:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Setting PP block ranks... +12/12/2024 17:29:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Total number of parameters: 1.04G (1984.26MiB) +12/12/2024 17:29:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Local number of parameters: 520M (992.13MiB) +12/12/2024 17:29:21 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-103]: Local number of parameters: 520M (992.13MiB) +12/12/2024 17:29:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/12/2024 17:29:21 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-103]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/12/2024 17:29:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: No checkpoint path provided. +12/12/2024 17:29:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Parametrizing model parameters using StandardParametrizator +12/12/2024 17:29:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Optimizer Building] Using LearningRateForSP as learning rate +12/12/2024 17:29:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] Size of optimizer params per rank: +12/12/2024 17:29:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 0 has 130M out of 520M (25.00%) params' optimizer states +12/12/2024 17:29:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 1 has 130M out of 520M (25.00%) params' optimizer states +12/12/2024 17:29:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 2 has 130M out of 520M (25.00%) params' optimizer states +12/12/2024 17:29:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 3 has 130M out of 520M (25.00%) params' optimizer states +12/12/2024 17:29:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/12/2024 17:29:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Using dummy data generator +12/12/2024 17:29:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Training Plan] There are 1 training stages +12/12/2024 17:29:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Stage Stable Training Stage] start from step 1 +12/12/2024 17:29:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: +12/12/2024 17:29:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Start training] datetime: 2024-12-12 17:29:26.356517 | mbs: 2 | grad_accum: 256 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/12/2024 17:29:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/12/2024 17:29:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 3472.48MiB. Peak allocated 5632.00MiB. Peak reserved: 18884.00MiB +12/12/2024 17:29:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 3572.67MiB. Peak allocated 6985.16MiB. Peak reserved: 7966.00MiB +12/12/2024 17:29:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 21.5K | tokens_per_sec: 195K | tokens_per_sec_per_gpu: 24.4K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 172 | hardware_tflops_per_gpu: 172 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.66G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.4G | hd_free_memory_tb: 243G +12/12/2024 17:29:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4588.85MiB. Peak allocated 6821.14MiB. Peak reserved: 9210.00MiB +slurmstepd: error: *** STEP 13385524.0 ON ip-26-0-160-103 CANCELLED AT 2024-12-12T17:29:56 DUE TO TIME LIMIT *** +slurmstepd: error: *** JOB 13385524 ON ip-26-0-160-103 CANCELLED AT 2024-12-12T17:29:56 DUE TO TIME LIMIT *** +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-12 17:29:56,045] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-12 17:29:56,046] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871450 closing signal SIGTERM +[2024-12-12 17:29:56,046] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871451 closing signal SIGTERM +[2024-12-12 17:29:56,046] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871452 closing signal SIGTERM +[2024-12-12 17:29:56,046] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871453 closing signal SIGTERM +[2024-12-12 17:29:56,046] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871454 closing signal SIGTERM +[2024-12-12 17:29:56,046] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871455 closing signal SIGTERM +[2024-12-12 17:29:56,050] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871456 closing signal SIGTERM +[2024-12-12 17:29:56,050] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871457 closing signal SIGTERM +[2024-12-12 17:29:56,191] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871450 closing signal SIGTERM +[2024-12-12 17:29:56,191] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871451 closing signal SIGTERM +[2024-12-12 17:29:56,191] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871452 closing signal SIGTERM +[2024-12-12 17:29:56,191] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871453 closing signal SIGTERM +[2024-12-12 17:29:56,192] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871454 closing signal SIGTERM +[2024-12-12 17:29:56,192] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871455 closing signal SIGTERM +[2024-12-12 17:29:56,192] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871456 closing signal SIGTERM +[2024-12-12 17:29:56,192] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1871457 closing signal SIGTERM diff --git a/logs/13401593-bench_stress_test.out b/logs/13401593-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..683c5b1ef6cecae232ec9150bd3f04596af5a2ae --- /dev/null +++ b/logs/13401593-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-160-103 ++ export NODELIST=ip-26-0-160-103 ++ NODELIST=ip-26-0-160-103 +++ scontrol show hostnames ip-26-0-160-103 +++ head -n1 ++ export MASTER_NODE=ip-26-0-160-103 ++ MASTER_NODE=ip-26-0-160-103 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-103' +Master node: ip-26-0-160-103 ++ echo 'All nodes: ip-26-0-160-103' +All nodes: ip-26-0-160-103 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13401593 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-103:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 00:02:36,819] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 00:02:36,819] torch.distributed.run: [WARNING] +[2024-12-13 00:02:36,819] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 00:02:36,819] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 00:02:36,819] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 00:02:41,950] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 2117357) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_00:02:41 + host : ip-26-0-160-103.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 2117358) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_00:02:41 + host : ip-26-0-160-103.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 2117359) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_00:02:41 + host : ip-26-0-160-103.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 2117360) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_00:02:41 + host : ip-26-0-160-103.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 2117361) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_00:02:41 + host : ip-26-0-160-103.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 2117362) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_00:02:41 + host : ip-26-0-160-103.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 2117363) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_00:02:41 + host : ip-26-0-160-103.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 2117364) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_00:02:41 + host : ip-26-0-160-103.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 2117357) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-160-103: task 0: Exited with exit code 1 diff --git a/logs/13401650-bench_stress_test.out b/logs/13401650-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..28c6ec0ab6383877394f2bcc39d3edd259d88a59 --- /dev/null +++ b/logs/13401650-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13401650 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 00:11:55,060] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 00:11:55,060] torch.distributed.run: [WARNING] +[2024-12-13 00:11:55,060] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 00:11:55,060] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 00:11:55,060] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 00:12:00,161] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 234324) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_00:12:00 + host : ip-26-0-169-207.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 234325) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_00:12:00 + host : ip-26-0-169-207.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 234326) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_00:12:00 + host : ip-26-0-169-207.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 234327) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_00:12:00 + host : ip-26-0-169-207.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 234328) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_00:12:00 + host : ip-26-0-169-207.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 234329) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_00:12:00 + host : ip-26-0-169-207.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 234330) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_00:12:00 + host : ip-26-0-169-207.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 234331) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_00:12:00 + host : ip-26-0-169-207.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 234324) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-169-207: task 0: Exited with exit code 1 diff --git a/logs/13401711-bench_stress_test.out b/logs/13401711-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..3b3616d3a6238c7ace70544b655adabb68f56cce --- /dev/null +++ b/logs/13401711-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13401711 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 00:21:55,762] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 00:21:55,762] torch.distributed.run: [WARNING] +[2024-12-13 00:21:55,762] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 00:21:55,762] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 00:21:55,762] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 00:22:01,009] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 243992) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_00:22:01 + host : ip-26-0-169-207.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 243993) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_00:22:01 + host : ip-26-0-169-207.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 243994) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_00:22:01 + host : ip-26-0-169-207.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 243995) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_00:22:01 + host : ip-26-0-169-207.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 243996) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_00:22:01 + host : ip-26-0-169-207.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 243997) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_00:22:01 + host : ip-26-0-169-207.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 243998) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_00:22:01 + host : ip-26-0-169-207.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 243999) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_00:22:01 + host : ip-26-0-169-207.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 243992) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-169-207: task 0: Exited with exit code 1 diff --git a/logs/13401821-bench_stress_test.out b/logs/13401821-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..d6438b5f61869de6cfd191f1fbf796c85e2715c8 --- /dev/null +++ b/logs/13401821-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13401821 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 00:40:18,675] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 00:40:18,675] torch.distributed.run: [WARNING] +[2024-12-13 00:40:18,675] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 00:40:18,675] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 00:40:18,675] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 00:40:23,900] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 261213) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_00:40:23 + host : ip-26-0-169-207.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 261214) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_00:40:23 + host : ip-26-0-169-207.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 261215) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_00:40:23 + host : ip-26-0-169-207.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 261216) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_00:40:23 + host : ip-26-0-169-207.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 261217) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_00:40:23 + host : ip-26-0-169-207.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 261218) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_00:40:23 + host : ip-26-0-169-207.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 261219) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_00:40:23 + host : ip-26-0-169-207.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 261220) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_00:40:23 + host : ip-26-0-169-207.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 261213) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-169-207: task 0: Exited with exit code 1 diff --git a/logs/13401960-bench_stress_test.out b/logs/13401960-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..699b98c1ea637a9e4e9352224734286aec755cc3 --- /dev/null +++ b/logs/13401960-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13401960 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 01:03:51,480] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 01:03:51,480] torch.distributed.run: [WARNING] +[2024-12-13 01:03:51,480] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 01:03:51,480] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 01:03:51,480] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 01:03:56,650] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 283585) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_01:03:56 + host : ip-26-0-169-207.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 283586) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_01:03:56 + host : ip-26-0-169-207.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 283587) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_01:03:56 + host : ip-26-0-169-207.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 283588) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_01:03:56 + host : ip-26-0-169-207.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 283589) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_01:03:56 + host : ip-26-0-169-207.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 283590) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_01:03:56 + host : ip-26-0-169-207.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 283591) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_01:03:56 + host : ip-26-0-169-207.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 283592) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_01:03:56 + host : ip-26-0-169-207.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 283585) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-169-207: task 0: Exited with exit code 1 diff --git a/logs/13402468-bench_stress_test.out b/logs/13402468-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..ff498a8c989167f5c8ad50091c8b89a0371c5a4a --- /dev/null +++ b/logs/13402468-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13402468 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 02:28:54,760] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 02:28:54,760] torch.distributed.run: [WARNING] +[2024-12-13 02:28:54,760] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 02:28:54,760] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 02:28:54,760] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 02:28:59,957] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 363461) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_02:28:59 + host : ip-26-0-169-207.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 363462) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_02:28:59 + host : ip-26-0-169-207.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 363463) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_02:28:59 + host : ip-26-0-169-207.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 363464) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_02:28:59 + host : ip-26-0-169-207.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 363465) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_02:28:59 + host : ip-26-0-169-207.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 363466) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_02:28:59 + host : ip-26-0-169-207.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 363467) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_02:28:59 + host : ip-26-0-169-207.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 363468) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_02:28:59 + host : ip-26-0-169-207.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 363461) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-169-207: task 0: Exited with exit code 1 diff --git a/logs/13402487-bench_stress_test.out b/logs/13402487-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..bff777b713df58a6771f9fd46acb70c75e8920ef --- /dev/null +++ b/logs/13402487-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-160-103 ++ export NODELIST=ip-26-0-160-103 ++ NODELIST=ip-26-0-160-103 +++ scontrol show hostnames ip-26-0-160-103 +++ head -n1 ++ export MASTER_NODE=ip-26-0-160-103 ++ MASTER_NODE=ip-26-0-160-103 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-103' +Master node: ip-26-0-160-103 ++ echo 'All nodes: ip-26-0-160-103' +All nodes: ip-26-0-160-103 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13402487 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-103:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 02:32:04,895] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 02:32:04,895] torch.distributed.run: [WARNING] +[2024-12-13 02:32:04,895] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 02:32:04,895] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 02:32:04,895] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 02:32:10,217] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 2256812) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_02:32:10 + host : ip-26-0-160-103.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 2256813) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_02:32:10 + host : ip-26-0-160-103.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 2256814) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_02:32:10 + host : ip-26-0-160-103.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 2256815) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_02:32:10 + host : ip-26-0-160-103.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 2256816) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_02:32:10 + host : ip-26-0-160-103.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 2256817) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_02:32:10 + host : ip-26-0-160-103.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 2256818) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_02:32:10 + host : ip-26-0-160-103.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 2256819) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_02:32:10 + host : ip-26-0-160-103.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 2256812) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-160-103: task 0: Exited with exit code 1 diff --git a/logs/13404213-bench_stress_test.out b/logs/13404213-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..9aae8f929b4e001cd2babed84d821f99443cbb39 --- /dev/null +++ b/logs/13404213-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-160-103 ++ export NODELIST=ip-26-0-160-103 ++ NODELIST=ip-26-0-160-103 +++ scontrol show hostnames ip-26-0-160-103 +++ head -n1 ++ export MASTER_NODE=ip-26-0-160-103 ++ MASTER_NODE=ip-26-0-160-103 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-103' +Master node: ip-26-0-160-103 ++ echo 'All nodes: ip-26-0-160-103' +All nodes: ip-26-0-160-103 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13404213 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-103:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 03:11:19,299] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 03:11:19,299] torch.distributed.run: [WARNING] +[2024-12-13 03:11:19,299] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 03:11:19,299] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 03:11:19,299] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 03:11:24,415] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 2295225) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_03:11:24 + host : ip-26-0-160-103.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 2295226) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_03:11:24 + host : ip-26-0-160-103.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 2295227) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_03:11:24 + host : ip-26-0-160-103.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 2295228) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_03:11:24 + host : ip-26-0-160-103.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 2295229) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_03:11:24 + host : ip-26-0-160-103.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 2295230) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_03:11:24 + host : ip-26-0-160-103.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 2295231) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_03:11:24 + host : ip-26-0-160-103.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 2295232) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_03:11:24 + host : ip-26-0-160-103.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 2295225) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-160-103: task 0: Exited with exit code 1 diff --git a/logs/13404445-bench_stress_test.out b/logs/13404445-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..87e52dab5a511dc3a19a41f525972e4555bb3674 --- /dev/null +++ b/logs/13404445-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13404445 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 03:47:20,106] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 03:47:20,106] torch.distributed.run: [WARNING] +[2024-12-13 03:47:20,106] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 03:47:20,106] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 03:47:20,106] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 03:47:25,381] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 441404) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_03:47:25 + host : ip-26-0-169-207.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 441405) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_03:47:25 + host : ip-26-0-169-207.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 441406) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_03:47:25 + host : ip-26-0-169-207.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 441407) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_03:47:25 + host : ip-26-0-169-207.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 441408) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_03:47:25 + host : ip-26-0-169-207.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 441409) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_03:47:25 + host : ip-26-0-169-207.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 441410) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_03:47:25 + host : ip-26-0-169-207.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 441411) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_03:47:25 + host : ip-26-0-169-207.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 441404) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-169-207: task 0: Exited with exit code 1 diff --git a/logs/13404677-bench_stress_test.out b/logs/13404677-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..fdf067d40cd187e418ff039ed9e8a16eee7ac1ce --- /dev/null +++ b/logs/13404677-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13404677 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 04:23:36,713] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 04:23:36,713] torch.distributed.run: [WARNING] +[2024-12-13 04:23:36,713] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 04:23:36,713] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 04:23:36,713] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 04:23:41,958] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 477865) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_04:23:41 + host : ip-26-0-169-207.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 477866) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_04:23:41 + host : ip-26-0-169-207.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 477867) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_04:23:41 + host : ip-26-0-169-207.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 477868) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_04:23:41 + host : ip-26-0-169-207.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 477869) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_04:23:41 + host : ip-26-0-169-207.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 477870) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_04:23:41 + host : ip-26-0-169-207.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 477871) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_04:23:41 + host : ip-26-0-169-207.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 477872) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_04:23:41 + host : ip-26-0-169-207.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 477865) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-169-207: task 0: Exited with exit code 1 diff --git a/logs/13404698-bench_stress_test.out b/logs/13404698-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..c22d8d8d8428c7ad119bea9361803ffd5a39829a --- /dev/null +++ b/logs/13404698-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-160-103 ++ export NODELIST=ip-26-0-160-103 ++ NODELIST=ip-26-0-160-103 +++ scontrol show hostnames ip-26-0-160-103 +++ head -n1 ++ export MASTER_NODE=ip-26-0-160-103 ++ MASTER_NODE=ip-26-0-160-103 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-103' +Master node: ip-26-0-160-103 ++ echo 'All nodes: ip-26-0-160-103' +All nodes: ip-26-0-160-103 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13404698 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-103:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 04:27:15,286] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 04:27:15,286] torch.distributed.run: [WARNING] +[2024-12-13 04:27:15,286] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 04:27:15,286] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 04:27:15,286] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 04:27:20,563] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 2370383) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_04:27:20 + host : ip-26-0-160-103.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 2370384) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_04:27:20 + host : ip-26-0-160-103.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 2370385) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_04:27:20 + host : ip-26-0-160-103.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 2370386) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_04:27:20 + host : ip-26-0-160-103.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 2370387) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_04:27:20 + host : ip-26-0-160-103.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 2370388) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_04:27:20 + host : ip-26-0-160-103.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 2370389) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_04:27:20 + host : ip-26-0-160-103.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 2370390) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_04:27:20 + host : ip-26-0-160-103.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 2370383) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-160-103: task 0: Exited with exit code 1 diff --git a/logs/13404947-bench_stress_test.out b/logs/13404947-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..ed46f485d20af1684238d52a8fd5cb0679bc1dd7 --- /dev/null +++ b/logs/13404947-bench_stress_test.out @@ -0,0 +1,518 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-160-103 ++ export NODELIST=ip-26-0-160-103 ++ NODELIST=ip-26-0-160-103 +++ scontrol show hostnames ip-26-0-160-103 +++ head -n1 ++ export MASTER_NODE=ip-26-0-160-103 ++ MASTER_NODE=ip-26-0-160-103 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-103' +Master node: ip-26-0-160-103 ++ echo 'All nodes: ip-26-0-160-103' +All nodes: ip-26-0-160-103 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13404947 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-103:12356 run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 04:58:25,166] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 04:58:25,166] torch.distributed.run: [WARNING] +[2024-12-13 04:58:25,166] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 04:58:25,166] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 04:58:25,166] torch.distributed.run: [WARNING] ***************************************** +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +/fsx/nouamane/miniconda/envs/2-1-cu121/bin/python: can't open file '/fsx/nouamane/projects/run_train.py': [Errno 2] No such file or directory +[2024-12-13 04:58:30,501] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 2) local_rank: 0 (pid: 2403960) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-13_04:58:30 + host : ip-26-0-160-103.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 2 (pid: 2403961) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-13_04:58:30 + host : ip-26-0-160-103.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 2 (pid: 2403962) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-13_04:58:30 + host : ip-26-0-160-103.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 2 (pid: 2403963) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-13_04:58:30 + host : ip-26-0-160-103.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 2 (pid: 2403964) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-13_04:58:30 + host : ip-26-0-160-103.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 2 (pid: 2403965) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-13_04:58:30 + host : ip-26-0-160-103.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 2 (pid: 2403966) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-13_04:58:30 + host : ip-26-0-160-103.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 2 (pid: 2403967) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-13_04:58:30 + host : ip-26-0-160-103.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 2 (pid: 2403960) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-160-103: task 0: Exited with exit code 1 diff --git a/logs/13417106-bench_stress_test.out b/logs/13417106-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..451f149760a9ac2d66177955fb2fec6e3a1366ba --- /dev/null +++ b/logs/13417106-bench_stress_test.out @@ -0,0 +1,4106 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-160-103 ++ export NODELIST=ip-26-0-160-103 ++ NODELIST=ip-26-0-160-103 +++ scontrol show hostnames ip-26-0-160-103 +++ head -n1 ++ export MASTER_NODE=ip-26-0-160-103 ++ MASTER_NODE=ip-26-0-160-103 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-103' +Master node: ip-26-0-160-103 ++ echo 'All nodes: ip-26-0-160-103' +All nodes: ip-26-0-160-103 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13417106 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-103:12356 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 09:10:25,284] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 09:10:25,284] torch.distributed.run: [WARNING] +[2024-12-13 09:10:25,284] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 09:10:25,284] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 09:10:25,284] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-160-103:2479992:2479992 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2479992:2479992 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2479992:2479992 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2479992:2479992 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2479992:2479992 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2479993:2479993 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2479996:2479996 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2479993:2479993 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2479996:2479996 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2479993:2479993 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2479996:2479996 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2479993:2479993 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2479993:2479993 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2479996:2479996 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2479996:2479996 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2479998:2479998 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2479998:2479998 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2479998:2479998 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2479998:2479998 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2479998:2479998 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2479997:2479997 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2479997:2479997 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2479994:2479994 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2479997:2479997 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2479999:2479999 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2479994:2479994 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2479999:2479999 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2479997:2479997 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2479997:2479997 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2479994:2479994 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2479999:2479999 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2479994:2479994 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2479994:2479994 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2479999:2479999 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2479999:2479999 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2479995:2479995 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2479995:2479995 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2479995:2479995 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2479995:2479995 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2479995:2479995 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO comm 0x9977d30 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xafdd12b6fa14ff8c - Init START +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO comm 0x8d72bd0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xafdd12b6fa14ff8c - Init START +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO comm 0x8e650d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xafdd12b6fa14ff8c - Init START +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO comm 0x95e59f0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xafdd12b6fa14ff8c - Init START +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO comm 0x926c040 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xafdd12b6fa14ff8c - Init START +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO comm 0x8fc4820 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xafdd12b6fa14ff8c - Init START +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO comm 0x8baf950 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xafdd12b6fa14ff8c - Init START +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO comm 0x98e1370 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xafdd12b6fa14ff8c - Init START +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NVLS comm 0x98e1370 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NVLS comm 0x926c040 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NVLS comm 0x8fc4820 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NVLS comm 0x95e59f0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NVLS comm 0x8d72bd0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NVLS comm 0x8baf950 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NVLS comm 0x8e650d0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NVLS comm 0x9977d30 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479997:2480119 [5] NCCL INFO comm 0x8baf950 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xafdd12b6fa14ff8c - Init COMPLETE +ip-26-0-160-103:2479995:2480120 [3] NCCL INFO comm 0x8d72bd0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xafdd12b6fa14ff8c - Init COMPLETE +ip-26-0-160-103:2479999:2480118 [7] NCCL INFO comm 0x926c040 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xafdd12b6fa14ff8c - Init COMPLETE +ip-26-0-160-103:2479996:2480114 [4] NCCL INFO comm 0x9977d30 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xafdd12b6fa14ff8c - Init COMPLETE +ip-26-0-160-103:2479992:2480113 [0] NCCL INFO comm 0x98e1370 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xafdd12b6fa14ff8c - Init COMPLETE +ip-26-0-160-103:2479993:2480116 [1] NCCL INFO comm 0x8fc4820 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xafdd12b6fa14ff8c - Init COMPLETE +ip-26-0-160-103:2479998:2480115 [6] NCCL INFO comm 0x8e650d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xafdd12b6fa14ff8c - Init COMPLETE +ip-26-0-160-103:2479994:2480117 [2] NCCL INFO comm 0x95e59f0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xafdd12b6fa14ff8c - Init COMPLETE +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO comm 0x91edd10 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x24851fb62e8b4400 - Init START +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO comm 0x980f730 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x24851fb62e8b4400 - Init START +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO comm 0x908ee70 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x24851fb62e8b4400 - Init START +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO comm 0x9ba1a70 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x24851fb62e8b4400 - Init START +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO comm 0x9495910 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x24851fb62e8b4400 - Init START +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO comm 0x8dd9510 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x24851fb62e8b4400 - Init START +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO comm 0x8f9c8f0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x24851fb62e8b4400 - Init START +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO comm 0x9b0ae70 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x24851fb62e8b4400 - Init START +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NVLS comm 0x9b0ae70 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NVLS comm 0x9495910 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NVLS comm 0x91edd10 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NVLS comm 0x8dd9510 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NVLS comm 0x8f9c8f0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NVLS comm 0x908ee70 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NVLS comm 0x980f730 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NVLS comm 0x9ba1a70 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479992:2480184 [0] NCCL INFO comm 0x9b0ae70 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x24851fb62e8b4400 - Init COMPLETE +ip-26-0-160-103:2479996:2480189 [4] NCCL INFO comm 0x9ba1a70 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x24851fb62e8b4400 - Init COMPLETE +ip-26-0-160-103:2479998:2480187 [6] NCCL INFO comm 0x908ee70 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x24851fb62e8b4400 - Init COMPLETE +ip-26-0-160-103:2479994:2480190 [2] NCCL INFO comm 0x980f730 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x24851fb62e8b4400 - Init COMPLETE +ip-26-0-160-103:2479993:2480191 [1] NCCL INFO comm 0x91edd10 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x24851fb62e8b4400 - Init COMPLETE +ip-26-0-160-103:2479997:2480188 [5] NCCL INFO comm 0x8dd9510 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x24851fb62e8b4400 - Init COMPLETE +ip-26-0-160-103:2479995:2480185 [3] NCCL INFO comm 0x8f9c8f0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x24851fb62e8b4400 - Init COMPLETE +ip-26-0-160-103:2479999:2480186 [7] NCCL INFO comm 0x9495910 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x24851fb62e8b4400 - Init COMPLETE +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO comm 0x9823e90 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7f764e32648de0b6 - Init START +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO comm 0x90a35d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7f764e32648de0b6 - Init START +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO comm 0x94a9fa0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7f764e32648de0b6 - Init START +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO comm 0x92022b0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7f764e32648de0b6 - Init START +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO comm 0x9bb61d0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7f764e32648de0b6 - Init START +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO comm 0x8dedc70 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7f764e32648de0b6 - Init START +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO comm 0x9b1f4b0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7f764e32648de0b6 - Init START +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO comm 0x8fb0e60 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7f764e32648de0b6 - Init START +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NVLS comm 0x9b1f4b0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NVLS comm 0x94a9fa0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NVLS comm 0x92022b0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NVLS comm 0x9823e90 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NVLS comm 0x90a35d0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NVLS comm 0x8fb0e60 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NVLS comm 0x8dedc70 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NVLS comm 0x9bb61d0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479997:2480225 [5] NCCL INFO comm 0x8dedc70 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7f764e32648de0b6 - Init COMPLETE +ip-26-0-160-103:2479996:2480219 [4] NCCL INFO comm 0x9bb61d0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7f764e32648de0b6 - Init COMPLETE +ip-26-0-160-103:2479995:2480220 [3] NCCL INFO comm 0x8fb0e60 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7f764e32648de0b6 - Init COMPLETE +ip-26-0-160-103:2479993:2480224 [1] NCCL INFO comm 0x92022b0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7f764e32648de0b6 - Init COMPLETE +ip-26-0-160-103:2479992:2480218 [0] NCCL INFO comm 0x9b1f4b0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7f764e32648de0b6 - Init COMPLETE +ip-26-0-160-103:2479999:2480223 [7] NCCL INFO comm 0x94a9fa0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7f764e32648de0b6 - Init COMPLETE +ip-26-0-160-103:2479998:2480222 [6] NCCL INFO comm 0x90a35d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7f764e32648de0b6 - Init COMPLETE +ip-26-0-160-103:2479994:2480221 [2] NCCL INFO comm 0x9823e90 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7f764e32648de0b6 - Init COMPLETE +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Config: +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Config(general=GeneralArgs(project='debug', +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: run='1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k', +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: seed=42, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: step=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: consumed_train_samples=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: ignore_sanity_checks=True), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: parallelism=ParallelismArgs(dp=4, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pp=1, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp=2, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pp_engine=, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_mode=, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_linear_async_communication=True, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: recompute_layer=False, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_recompute_allgather=True, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: expert_parallel_size=1), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: eos_token_id=0, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_act='silu', +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_size=2048, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: initializer_range=0.02, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: intermediate_size=8192, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: is_llama_config=True, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: max_position_embeddings=2048, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_attention_heads=32, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_hidden_layers=16, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_key_value_heads=8, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pad_token_id=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pretraining_tp=1, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rms_norm_eps=1e-05, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_scaling=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_theta=10000.0, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_interleaved=False, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tie_word_embeddings=True, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: use_cache=True, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: vocab_size=32768), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: init_method=RandomInit(std=0.02), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: dtype=torch.bfloat16, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: make_vocab_size_divisible_by=1, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: ddp_bucket_cap_mb=25), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer_revision=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer_max_length=None), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoint_interval=10000, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: save_initial_state=False, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: save_final_state=False, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: resume_checkpoint_path=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoints_path_is_shared_file_system=False), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: logging=LoggingArgs(log_level='info', +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: log_level_replica='info', +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration_step_info_interval=1), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokens=TokensArgs(sequence_length=2048, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: train_steps=100, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: micro_batch_size=2, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: batch_accumulation_per_replica=256, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: val_check_interval=100, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: limit_val_batches=0, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: limit_test_batches=0), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: adam_beta1=0.9, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: adam_beta2=0.95, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: torch_adam_is_fused=True, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: name='adamW'), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: zero_stage=1, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: weight_decay=0.01, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: clip_grad=1.0, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: accumulate_grad_in_fp32=True, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_warmup_steps=2, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_warmup_style='linear', +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_style='cosine', +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_steps=13, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_starting_step=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: min_decay_lr=1e-05)), +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: start_training_step=1, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: data=DataArgs(dataset=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: seed=42, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_loading_workers=1))], +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: profiler=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lighteval=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: s3_upload=None) +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Model Config: +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: LlamaConfig(bos_token_id=0, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: eos_token_id=0, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_act='silu', +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_size=2048, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: initializer_range=0.02, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: intermediate_size=8192, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: is_llama_config=True, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: max_position_embeddings=2048, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_attention_heads=32, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_hidden_layers=16, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_key_value_heads=8, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pad_token_id=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pretraining_tp=1, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rms_norm_eps=1e-05, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_scaling=None, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_theta=10000.0, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_interleaved=False, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tie_word_embeddings=True, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: use_cache=True, +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: vocab_size=32768) +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Building model.. +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Initialize RoPE Theta = 10000.0 +12/13/2024 09:10:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO comm 0xb9b77b0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb1f03a39782ca6e8 - Init START +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO comm 0xabf1ab0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb1f03a39782ca6e8 - Init START +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO comm 0xaea7870 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7c55f716c66b1778 - Init START +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO comm 0xb2ab8a0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7c55f716c66b1778 - Init START +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO comm 0xb924d60 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xeba77532677df7d5 - Init START +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO comm 0xb002de0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeba77532677df7d5 - Init START +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO comm 0xb624e70 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe49768830ff674db - Init START +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO comm 0xadb4f60 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe49768830ff674db - Init START +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479992:2480272 [0] NCCL INFO comm 0xb924d60 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xeba77532677df7d5 - Init COMPLETE +ip-26-0-160-103:2479993:2480278 [1] NCCL INFO comm 0xb002de0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeba77532677df7d5 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO comm 0xb016a10 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4dbd53f8ad3c944c - Init START +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO comm 0xb938760 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb17e80ebd3e9643a - Init START +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479995:2480279 [3] NCCL INFO comm 0xadb4f60 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe49768830ff674db - Init COMPLETE +ip-26-0-160-103:2479994:2480275 [2] NCCL INFO comm 0xb624e70 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe49768830ff674db - Init COMPLETE +ip-26-0-160-103:2479998:2480274 [6] NCCL INFO comm 0xaea7870 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7c55f716c66b1778 - Init COMPLETE +ip-26-0-160-103:2479999:2480276 [7] NCCL INFO comm 0xb2ab8a0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7c55f716c66b1778 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO comm 0xb2bf0a0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xdf7e0adeaa0d55ef - Init START +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO comm 0xaebb4a0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbadb60aa4c18e8fb - Init START +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO comm 0xadc8870 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x63811bfec708d8ea - Init START +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO comm 0xb638ba0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x225e408891d2965 - Init START +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479997:2480277 [5] NCCL INFO comm 0xabf1ab0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb1f03a39782ca6e8 - Init COMPLETE +ip-26-0-160-103:2479996:2480271 [4] NCCL INFO comm 0xb9b77b0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb1f03a39782ca6e8 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO comm 0xac05700 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3be9e4c9ab0c2b1f - Init START +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO comm 0xb9cbb20 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd611a833a57ca2e8 - Init START +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480303 [7] NCCL INFO comm 0xb2bf0a0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xdf7e0adeaa0d55ef - Init COMPLETE +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479994:2480308 [2] NCCL INFO comm 0xb638ba0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x225e408891d2965 - Init COMPLETE +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480304 [6] NCCL INFO comm 0xaebb4a0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbadb60aa4c18e8fb - Init COMPLETE +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479995:2480306 [3] NCCL INFO comm 0xadc8870 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x63811bfec708d8ea - Init COMPLETE +ip-26-0-160-103:2479993:2480298 [1] NCCL INFO comm 0xb016a10 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4dbd53f8ad3c944c - Init COMPLETE +ip-26-0-160-103:2479992:2480300 [0] NCCL INFO comm 0xb938760 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb17e80ebd3e9643a - Init COMPLETE +12/13/2024 09:11:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Total number of parameters: 1.04G (1984.26MiB) +12/13/2024 09:11:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Local number of parameters: 520M (992.13MiB) +12/13/2024 09:11:03 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-103]: Local number of parameters: 520M (992.13MiB) +12/13/2024 09:11:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 17922.00MiB +12/13/2024 09:11:03 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-103]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/13/2024 09:11:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: No checkpoint path provided. +12/13/2024 09:11:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Parametrizing model parameters using StandardParametrizator +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479996:2480314 [4] NCCL INFO comm 0xb9cbb20 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd611a833a57ca2e8 - Init COMPLETE +ip-26-0-160-103:2479997:2480313 [5] NCCL INFO comm 0xac05700 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3be9e4c9ab0c2b1f - Init COMPLETE +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO comm 0xac0a790 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x91158aabee2f5d5 - Init START +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO comm 0xb01bbb0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x91158aabee2f5d5 - Init START +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO comm 0xb2c4060 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x91158aabee2f5d5 - Init START +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO comm 0xadcda90 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x91158aabee2f5d5 - Init START +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO comm 0xb9d0480 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x48a21c88d204720c - Init START +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO comm 0xb93dfc0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x48a21c88d204720c - Init START +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO comm 0xaec0870 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x48a21c88d204720c - Init START +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO comm 0xb63dd70 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x48a21c88d204720c - Init START +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 04/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 04/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 05/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 05/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 06/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 04/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 06/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 07/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 05/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 07/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 06/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 08/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 08/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 09/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 07/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 09/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 10/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 08/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 10/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 11/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 09/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 11/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 12/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 10/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 12/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 13/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 11/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 13/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 14/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 14/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 12/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 15/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 15/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 13/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 16/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 14/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 16/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 17/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 17/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 15/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 18/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 19/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 18/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 16/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 17/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 18/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 19/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 20/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 19/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 20/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 21/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 20/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 21/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 22/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 23/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 22/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 21/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 23/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 22/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 23/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 04/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 05/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 06/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 04/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 07/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 05/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 08/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 06/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 09/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 07/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 10/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 08/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 11/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 09/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 12/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 10/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 13/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 11/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 14/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 12/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 13/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 14/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 15/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 15/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 16/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 16/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 17/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 18/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 17/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 19/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 20/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 18/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 21/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 19/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 22/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 20/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 23/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 21/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 22/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 23/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 02/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 03/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 04/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 05/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 06/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 07/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 08/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 09/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 10/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 11/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 12/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 13/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 14/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 15/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 16/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 17/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 18/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 19/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 20/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 21/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 22/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Channel 23/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 04/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 05/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 06/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 07/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 08/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 09/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 10/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 11/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 12/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 13/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 14/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 15/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 16/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 17/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 18/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 19/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 20/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 21/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 22/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Channel 23/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 04/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 05/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 06/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 07/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 08/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 09/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NVLS comm 0xb93dfc0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 10/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NVLS comm 0xaec0870 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 11/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 12/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 13/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 14/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 15/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 16/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 17/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 18/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 19/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 20/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NVLS comm 0xb9d0480 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 21/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 22/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Channel 23/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NVLS comm 0xb63dd70 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479998:2480331 [6] NCCL INFO comm 0xaec0870 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x48a21c88d204720c - Init COMPLETE +ip-26-0-160-103:2479996:2480342 [4] NCCL INFO comm 0xb9d0480 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x48a21c88d204720c - Init COMPLETE +ip-26-0-160-103:2479992:2480330 [0] NCCL INFO comm 0xb93dfc0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x48a21c88d204720c - Init COMPLETE +ip-26-0-160-103:2479994:2480333 [2] NCCL INFO comm 0xb63dd70 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x48a21c88d204720c - Init COMPLETE +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NVLS comm 0xb2c4060 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NVLS comm 0xb01bbb0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NVLS comm 0xadcda90 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NVLS comm 0xac0a790 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2479995:2480336 [3] NCCL INFO comm 0xadcda90 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x91158aabee2f5d5 - Init COMPLETE +ip-26-0-160-103:2479999:2480335 [7] NCCL INFO comm 0xb2c4060 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x91158aabee2f5d5 - Init COMPLETE +ip-26-0-160-103:2479997:2480341 [5] NCCL INFO comm 0xac0a790 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x91158aabee2f5d5 - Init COMPLETE +ip-26-0-160-103:2479993:2480334 [1] NCCL INFO comm 0xb01bbb0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x91158aabee2f5d5 - Init COMPLETE +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Optimizer Building] Using LearningRateForSP as learning rate +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] Size of optimizer params per rank: +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 0 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 1 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 2 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 3 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Using dummy data generator +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Training Plan] There are 1 training stages +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Stage Stable Training Stage] start from step 1 +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: +12/13/2024 09:11:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Start training] datetime: 2024-12-13 09:11:07.861558 | mbs: 2 | grad_accum: 256 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/13/2024 09:11:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/13/2024 09:11:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Waiting for 5 seconds +12/13/2024 09:11:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 3472.48MiB. Peak allocated 5632.00MiB. Peak reserved: 19908.00MiB +slurmstepd: error: *** JOB 13417106 ON ip-26-0-160-103 CANCELLED AT 2024-12-13T09:11:42 DUE TO TIME LIMIT *** +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +slurmstepd: error: *** STEP 13417106.0 ON ip-26-0-160-103 CANCELLED AT 2024-12-13T09:11:42 DUE TO TIME LIMIT *** +[2024-12-13 09:11:42,642] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-13 09:11:42,643] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479992 closing signal SIGTERM +[2024-12-13 09:11:42,643] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479993 closing signal SIGTERM +[2024-12-13 09:11:42,643] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479994 closing signal SIGTERM +[2024-12-13 09:11:42,644] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479995 closing signal SIGTERM +[2024-12-13 09:11:42,644] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479996 closing signal SIGTERM +[2024-12-13 09:11:42,644] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479997 closing signal SIGTERM +[2024-12-13 09:11:42,644] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479998 closing signal SIGTERM +[2024-12-13 09:11:42,646] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479999 closing signal SIGTERM +[2024-12-13 09:11:42,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479992 closing signal SIGTERM +[2024-12-13 09:11:42,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479993 closing signal SIGTERM +[2024-12-13 09:11:42,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479994 closing signal SIGTERM +[2024-12-13 09:11:42,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479995 closing signal SIGTERM +[2024-12-13 09:11:42,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479996 closing signal SIGTERM +[2024-12-13 09:11:42,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479997 closing signal SIGTERM +[2024-12-13 09:11:42,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479998 closing signal SIGTERM +[2024-12-13 09:11:42,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2479999 closing signal SIGTERM diff --git a/logs/13417423-bench_stress_test.out b/logs/13417423-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..04d29fc0fd5feb77d4135c3588f954e3b6c2bd5a --- /dev/null +++ b/logs/13417423-bench_stress_test.out @@ -0,0 +1,3767 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13417423 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 10:17:03,718] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 10:17:03,718] torch.distributed.run: [WARNING] +[2024-12-13 10:17:03,718] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 10:17:03,718] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 10:17:03,718] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-169-207:619094:619094 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:619094:619094 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:619094:619094 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:619094:619094 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:619094:619094 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:619097:619097 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:619097:619097 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:619100:619100 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:619100:619100 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:619095:619095 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:619095:619095 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:619097:619097 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:619099:619099 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:619099:619099 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:619100:619100 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:619096:619096 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:619097:619097 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:619097:619097 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:619095:619095 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:619096:619096 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:619100:619100 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:619100:619100 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:619099:619099 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:619095:619095 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:619095:619095 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:619096:619096 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:619099:619099 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:619099:619099 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:619096:619096 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:619096:619096 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:619101:619101 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:619101:619101 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:619101:619101 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:619101:619101 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:619101:619101 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:619098:619098 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:619098:619098 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:619098:619098 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:619098:619098 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:619098:619098 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/162 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:619095:619217 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:619095:619217 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:619098:619221 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:619098:619221 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:619096:619218 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:619096:619218 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:619099:619219 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:619099:619219 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:619094:619216 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:619094:619216 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:619100:619220 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:619100:619220 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:619097:619223 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:619097:619223 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:619101:619222 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:619101:619222 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:619099:619219 [5] NCCL INFO comm 0x9ff2ac0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9360c38bd1e404ba - Init START +ip-26-0-169-207:619097:619223 [3] NCCL INFO comm 0x85827d0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9360c38bd1e404ba - Init START +ip-26-0-169-207:619100:619220 [6] NCCL INFO comm 0x8666720 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9360c38bd1e404ba - Init START +ip-26-0-169-207:619094:619216 [0] NCCL INFO comm 0x9f40470 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9360c38bd1e404ba - Init START +ip-26-0-169-207:619101:619222 [7] NCCL INFO comm 0x8a2a590 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9360c38bd1e404ba - Init START +ip-26-0-169-207:619096:619218 [2] NCCL INFO comm 0x8cc31c0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9360c38bd1e404ba - Init START +ip-26-0-169-207:619095:619217 [1] NCCL INFO comm 0x8d08f90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9360c38bd1e404ba - Init START +ip-26-0-169-207:619098:619221 [4] NCCL INFO comm 0xa2ce370 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9360c38bd1e404ba - Init START +ip-26-0-169-207:619098:619221 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619097:619223 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619095:619217 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619096:619218 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619094:619216 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:619098:619221 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619098:619221 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:619094:619216 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619099:619219 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:619096:619218 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:619096:619218 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:619095:619217 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:619095:619217 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:619097:619223 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:619097:619223 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:619101:619222 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619101:619222 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:619100:619220 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619100:619220 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:619100:619220 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:619100:619220 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619099:619219 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:619097:619223 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:619095:619217 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:619099:619219 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619097:619223 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619095:619217 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619096:619218 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619096:619218 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619101:619222 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619101:619222 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619098:619221 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619098:619221 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:619094:619216 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Connected all rings +ip-26-0-169-207:619097:619223 [3] NCCL INFO Connected all rings +ip-26-0-169-207:619099:619219 [5] NCCL INFO Connected all rings +ip-26-0-169-207:619096:619218 [2] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619222 [7] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619094:619216 [0] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619218 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619219 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619223 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619222 [7] NCCL INFO Connected all trees +ip-26-0-169-207:619101:619222 [7] NCCL INFO NVLS comm 0x8a2a590 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619221 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619217 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619220 [6] NCCL INFO Connected all trees +ip-26-0-169-207:619100:619220 [6] NCCL INFO NVLS comm 0x8666720 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619094:619216 [0] NCCL INFO Connected all trees +ip-26-0-169-207:619094:619216 [0] NCCL INFO NVLS comm 0x9f40470 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619099:619219 [5] NCCL INFO Connected all trees +ip-26-0-169-207:619099:619219 [5] NCCL INFO NVLS comm 0x9ff2ac0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619096:619218 [2] NCCL INFO Connected all trees +ip-26-0-169-207:619096:619218 [2] NCCL INFO NVLS comm 0x8cc31c0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619098:619221 [4] NCCL INFO Connected all trees +ip-26-0-169-207:619098:619221 [4] NCCL INFO NVLS comm 0xa2ce370 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619097:619223 [3] NCCL INFO Connected all trees +ip-26-0-169-207:619097:619223 [3] NCCL INFO NVLS comm 0x85827d0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619095:619217 [1] NCCL INFO Connected all trees +ip-26-0-169-207:619095:619217 [1] NCCL INFO NVLS comm 0x8d08f90 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619095:619217 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619095:619217 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619095:619217 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619095:619217 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619099:619219 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619099:619219 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619099:619219 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619099:619219 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619097:619223 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619097:619223 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619097:619223 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619097:619223 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619101:619222 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619101:619222 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619101:619222 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619101:619222 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619094:619216 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619094:619216 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619094:619216 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619094:619216 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619100:619220 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619098:619221 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619100:619220 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619096:619218 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619100:619220 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619100:619220 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619098:619221 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619098:619221 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619098:619221 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619096:619218 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619096:619218 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619096:619218 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619094:619216 [0] NCCL INFO comm 0x9f40470 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9360c38bd1e404ba - Init COMPLETE +ip-26-0-169-207:619100:619220 [6] NCCL INFO comm 0x8666720 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9360c38bd1e404ba - Init COMPLETE +ip-26-0-169-207:619095:619217 [1] NCCL INFO comm 0x8d08f90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9360c38bd1e404ba - Init COMPLETE +ip-26-0-169-207:619099:619219 [5] NCCL INFO comm 0x9ff2ac0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9360c38bd1e404ba - Init COMPLETE +ip-26-0-169-207:619096:619218 [2] NCCL INFO comm 0x8cc31c0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9360c38bd1e404ba - Init COMPLETE +ip-26-0-169-207:619097:619223 [3] NCCL INFO comm 0x85827d0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9360c38bd1e404ba - Init COMPLETE +ip-26-0-169-207:619098:619221 [4] NCCL INFO comm 0xa2ce370 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9360c38bd1e404ba - Init COMPLETE +ip-26-0-169-207:619101:619222 [7] NCCL INFO comm 0x8a2a590 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9360c38bd1e404ba - Init COMPLETE +ip-26-0-169-207:619098:619291 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:619094:619287 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:619100:619288 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:619095:619290 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:619099:619292 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:619096:619293 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:619101:619289 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:619097:619294 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:619096:619293 [2] NCCL INFO comm 0x8eeccf0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xae6c3ce1561c19a - Init START +ip-26-0-169-207:619095:619290 [1] NCCL INFO comm 0x8f32690 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xae6c3ce1561c19a - Init START +ip-26-0-169-207:619094:619287 [0] NCCL INFO comm 0xa16a660 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xae6c3ce1561c19a - Init START +ip-26-0-169-207:619101:619289 [7] NCCL INFO comm 0x8c54610 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xae6c3ce1561c19a - Init START +ip-26-0-169-207:619097:619294 [3] NCCL INFO comm 0x87ac4d0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xae6c3ce1561c19a - Init START +ip-26-0-169-207:619100:619288 [6] NCCL INFO comm 0x8890480 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xae6c3ce1561c19a - Init START +ip-26-0-169-207:619099:619292 [5] NCCL INFO comm 0xa21c9d0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xae6c3ce1561c19a - Init START +ip-26-0-169-207:619098:619291 [4] NCCL INFO comm 0xa4f8000 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xae6c3ce1561c19a - Init START +ip-26-0-169-207:619095:619290 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619099:619292 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619094:619287 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:619097:619294 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619100:619288 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619098:619291 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619096:619293 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619101:619289 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619099:619292 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619099:619292 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:619096:619293 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:619096:619293 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:619095:619290 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:619095:619290 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:619094:619287 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:619098:619291 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619098:619291 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:619097:619294 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:619097:619294 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:619100:619288 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619100:619288 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:619101:619289 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619101:619289 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:619101:619289 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:619101:619289 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619097:619294 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:619100:619288 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:619098:619291 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:619097:619294 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619096:619293 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:619099:619292 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:619100:619288 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619098:619291 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619096:619293 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619099:619292 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619095:619290 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619287 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:619095:619290 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619094:619287 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Connected all rings +ip-26-0-169-207:619098:619291 [4] NCCL INFO Connected all rings +ip-26-0-169-207:619096:619293 [2] NCCL INFO Connected all rings +ip-26-0-169-207:619097:619294 [3] NCCL INFO Connected all rings +ip-26-0-169-207:619099:619292 [5] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619289 [7] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619290 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619291 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619292 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619293 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO Connected all trees +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619287 [0] NCCL INFO NVLS comm 0xa16a660 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619288 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619097:619294 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619101:619289 [7] NCCL INFO Connected all trees +ip-26-0-169-207:619101:619289 [7] NCCL INFO NVLS comm 0x8c54610 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619095:619290 [1] NCCL INFO Connected all trees +ip-26-0-169-207:619095:619290 [1] NCCL INFO NVLS comm 0x8f32690 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619096:619293 [2] NCCL INFO Connected all trees +ip-26-0-169-207:619096:619293 [2] NCCL INFO NVLS comm 0x8eeccf0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619098:619291 [4] NCCL INFO Connected all trees +ip-26-0-169-207:619098:619291 [4] NCCL INFO NVLS comm 0xa4f8000 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619100:619288 [6] NCCL INFO Connected all trees +ip-26-0-169-207:619100:619288 [6] NCCL INFO NVLS comm 0x8890480 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619099:619292 [5] NCCL INFO Connected all trees +ip-26-0-169-207:619099:619292 [5] NCCL INFO NVLS comm 0xa21c9d0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619097:619294 [3] NCCL INFO Connected all trees +ip-26-0-169-207:619097:619294 [3] NCCL INFO NVLS comm 0x87ac4d0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619099:619292 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619099:619292 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619099:619292 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619099:619292 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619101:619289 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619101:619289 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619101:619289 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619101:619289 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619097:619294 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619097:619294 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619100:619288 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619097:619294 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619097:619294 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619100:619288 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619100:619288 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619100:619288 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619098:619291 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619098:619291 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619098:619291 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619098:619291 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619096:619293 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619095:619290 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619096:619293 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619096:619293 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619096:619293 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619095:619290 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619095:619290 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619095:619290 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619094:619287 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619094:619287 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619094:619287 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619094:619287 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619098:619291 [4] NCCL INFO comm 0xa4f8000 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xae6c3ce1561c19a - Init COMPLETE +ip-26-0-169-207:619101:619289 [7] NCCL INFO comm 0x8c54610 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xae6c3ce1561c19a - Init COMPLETE +ip-26-0-169-207:619095:619290 [1] NCCL INFO comm 0x8f32690 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xae6c3ce1561c19a - Init COMPLETE +ip-26-0-169-207:619094:619287 [0] NCCL INFO comm 0xa16a660 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xae6c3ce1561c19a - Init COMPLETE +ip-26-0-169-207:619100:619288 [6] NCCL INFO comm 0x8890480 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xae6c3ce1561c19a - Init COMPLETE +ip-26-0-169-207:619096:619293 [2] NCCL INFO comm 0x8eeccf0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xae6c3ce1561c19a - Init COMPLETE +ip-26-0-169-207:619099:619292 [5] NCCL INFO comm 0xa21c9d0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xae6c3ce1561c19a - Init COMPLETE +ip-26-0-169-207:619097:619294 [3] NCCL INFO comm 0x87ac4d0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xae6c3ce1561c19a - Init COMPLETE +ip-26-0-169-207:619098:619322 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:619096:619326 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:619094:619320 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:619101:619321 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:619100:619323 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:619099:619324 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:619095:619325 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:619097:619327 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:619096:619326 [2] NCCL INFO comm 0x8f01450 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x924eaf138cc2c0e9 - Init START +ip-26-0-169-207:619095:619325 [1] NCCL INFO comm 0x8f46df0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x924eaf138cc2c0e9 - Init START +ip-26-0-169-207:619094:619320 [0] NCCL INFO comm 0xa17e9c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x924eaf138cc2c0e9 - Init START +ip-26-0-169-207:619100:619323 [6] NCCL INFO comm 0x88a4c20 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x924eaf138cc2c0e9 - Init START +ip-26-0-169-207:619101:619321 [7] NCCL INFO comm 0x8c68b20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x924eaf138cc2c0e9 - Init START +ip-26-0-169-207:619098:619322 [4] NCCL INFO comm 0xa50c770 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x924eaf138cc2c0e9 - Init START +ip-26-0-169-207:619097:619327 [3] NCCL INFO comm 0x87c0c40 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x924eaf138cc2c0e9 - Init START +ip-26-0-169-207:619099:619324 [5] NCCL INFO comm 0xa231130 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x924eaf138cc2c0e9 - Init START +ip-26-0-169-207:619095:619325 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619101:619321 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619096:619326 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619100:619323 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619099:619324 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619097:619327 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619098:619322 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619094:619320 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:619094:619320 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:619099:619324 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619099:619324 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:619098:619322 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619098:619322 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:619097:619327 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:619097:619327 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:619100:619323 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619100:619323 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:619101:619321 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619101:619321 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:619096:619326 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:619096:619326 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:619095:619325 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:619095:619325 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:619096:619326 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:619095:619325 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:619096:619326 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619095:619325 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619101:619321 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:619100:619323 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:619097:619327 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:619099:619324 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:619098:619322 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:619100:619323 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619101:619321 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619097:619327 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619099:619324 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619098:619322 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:619094:619320 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:619094:619320 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Connected all rings +ip-26-0-169-207:619095:619325 [1] NCCL INFO Connected all rings +ip-26-0-169-207:619097:619327 [3] NCCL INFO Connected all rings +ip-26-0-169-207:619094:619320 [0] NCCL INFO Connected all rings +ip-26-0-169-207:619099:619324 [5] NCCL INFO Connected all rings +ip-26-0-169-207:619100:619323 [6] NCCL INFO Connected all rings +ip-26-0-169-207:619096:619326 [2] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619321 [7] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:619098:619322 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619325 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619099:619324 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:619097:619327 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619096:619326 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619094:619320 [0] NCCL INFO Connected all trees +ip-26-0-169-207:619094:619320 [0] NCCL INFO NVLS comm 0xa17e9c0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619100:619323 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:619101:619321 [7] NCCL INFO Connected all trees +ip-26-0-169-207:619101:619321 [7] NCCL INFO NVLS comm 0x8c68b20 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619098:619322 [4] NCCL INFO Connected all trees +ip-26-0-169-207:619098:619322 [4] NCCL INFO NVLS comm 0xa50c770 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619100:619323 [6] NCCL INFO Connected all trees +ip-26-0-169-207:619100:619323 [6] NCCL INFO NVLS comm 0x88a4c20 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619096:619326 [2] NCCL INFO Connected all trees +ip-26-0-169-207:619096:619326 [2] NCCL INFO NVLS comm 0x8f01450 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619097:619327 [3] NCCL INFO Connected all trees +ip-26-0-169-207:619097:619327 [3] NCCL INFO NVLS comm 0x87c0c40 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619099:619324 [5] NCCL INFO Connected all trees +ip-26-0-169-207:619099:619324 [5] NCCL INFO NVLS comm 0xa231130 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619095:619325 [1] NCCL INFO Connected all trees +ip-26-0-169-207:619095:619325 [1] NCCL INFO NVLS comm 0x8f46df0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:619099:619324 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619099:619324 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619099:619324 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619099:619324 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619097:619327 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619101:619321 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619101:619321 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619097:619327 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619101:619321 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619101:619321 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619097:619327 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619097:619327 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619095:619325 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619095:619325 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619095:619325 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619095:619325 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619098:619322 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619100:619323 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619098:619322 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619100:619323 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619098:619322 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619100:619323 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619098:619322 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619100:619323 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619096:619326 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619096:619326 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619096:619326 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619096:619326 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619094:619320 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619094:619320 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619094:619320 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:619094:619320 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619100:619323 [6] NCCL INFO comm 0x88a4c20 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x924eaf138cc2c0e9 - Init COMPLETE +ip-26-0-169-207:619096:619326 [2] NCCL INFO comm 0x8f01450 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x924eaf138cc2c0e9 - Init COMPLETE +ip-26-0-169-207:619097:619327 [3] NCCL INFO comm 0x87c0c40 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x924eaf138cc2c0e9 - Init COMPLETE +ip-26-0-169-207:619101:619321 [7] NCCL INFO comm 0x8c68b20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x924eaf138cc2c0e9 - Init COMPLETE +ip-26-0-169-207:619099:619324 [5] NCCL INFO comm 0xa231130 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x924eaf138cc2c0e9 - Init COMPLETE +ip-26-0-169-207:619095:619325 [1] NCCL INFO comm 0x8f46df0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x924eaf138cc2c0e9 - Init COMPLETE +ip-26-0-169-207:619098:619322 [4] NCCL INFO comm 0xa50c770 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x924eaf138cc2c0e9 - Init COMPLETE +ip-26-0-169-207:619094:619320 [0] NCCL INFO comm 0xa17e9c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x924eaf138cc2c0e9 - Init COMPLETE +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config: +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config(general=GeneralArgs(project='debug', +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: run='1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k', +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: step=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: consumed_train_samples=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ignore_sanity_checks=True), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: parallelism=ParallelismArgs(dp=4, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp=1, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp=2, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp_engine=, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_mode=, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_linear_async_communication=True, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: recompute_layer=False, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_recompute_allgather=True, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: expert_parallel_size=1), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: init_method=RandomInit(std=0.02), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: dtype=torch.bfloat16, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: make_vocab_size_divisible_by=1, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ddp_bucket_cap_mb=25), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_revision=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_max_length=None), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoint_interval=10000, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_initial_state=False, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_final_state=False, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: resume_checkpoint_path=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints_path_is_shared_file_system=False), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: logging=LoggingArgs(log_level='info', +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: log_level_replica='info', +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration_step_info_interval=1), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokens=TokensArgs(sequence_length=2048, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: train_steps=100, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: micro_batch_size=2, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: batch_accumulation_per_replica=256, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: val_check_interval=100, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_val_batches=0, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_test_batches=0), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta1=0.9, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta2=0.95, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: torch_adam_is_fused=True, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: name='adamW'), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: zero_stage=1, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: weight_decay=0.01, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: clip_grad=1.0, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: accumulate_grad_in_fp32=True, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_steps=2, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_style='linear', +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_style='cosine', +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_steps=13, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_starting_step=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: min_decay_lr=1e-05)), +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: start_training_step=1, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data=DataArgs(dataset=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_loading_workers=1))], +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: profiler=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lighteval=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: s3_upload=None) +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Model Config: +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: LlamaConfig(bos_token_id=0, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768) +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Building model.. +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Initialize RoPE Theta = 10000.0 +12/13/2024 10:17:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:619101:619377 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:619098:619376 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:619100:619375 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:619096:619380 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:619094:619378 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:619099:619379 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:619095:619381 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:619097:619382 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:619101:619377 [7] NCCL INFO comm 0xaa6c220 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x941a314f46f8864f - Init START +ip-26-0-169-207:619100:619375 [6] NCCL INFO comm 0xa6a6860 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x941a314f46f8864f - Init START +ip-26-0-169-207:619100:619375 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619098:619376 [4] NCCL INFO comm 0xc30d5f0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x59df0f41f787b262 - Init START +ip-26-0-169-207:619099:619379 [5] NCCL INFO comm 0xc031670 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x59df0f41f787b262 - Init START +ip-26-0-169-207:619094:619378 [0] NCCL INFO comm 0xbf80f70 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc81bd109e66396f7 - Init START +ip-26-0-169-207:619095:619381 [1] NCCL INFO comm 0xad4a5f0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc81bd109e66396f7 - Init START +ip-26-0-169-207:619098:619376 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619094:619378 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:619096:619380 [2] NCCL INFO comm 0xad05d30 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44a96a70fdbb664d - Init START +ip-26-0-169-207:619097:619382 [3] NCCL INFO comm 0xa5c3cb0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44a96a70fdbb664d - Init START +ip-26-0-169-207:619096:619380 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619101:619377 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619099:619379 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619095:619381 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619097:619382 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619095:619381 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:619094:619378 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:619095:619381 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:619095:619381 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:619094:619378 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619097:619382 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:619098:619376 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619099:619379 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:619099:619379 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:619099:619379 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:619098:619376 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:619097:619382 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:619097:619382 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:619096:619380 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619101:619377 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:619101:619377 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:619101:619377 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:619100:619375 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619379 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619094:619378 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619098:619376 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619096:619380 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619097:619382 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619100:619375 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619101:619377 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:619095:619381 [1] NCCL INFO Connected all rings +ip-26-0-169-207:619095:619381 [1] NCCL INFO Connected all trees +ip-26-0-169-207:619095:619381 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619095:619381 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:619095:619381 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619094:619378 [0] NCCL INFO Connected all rings +ip-26-0-169-207:619094:619378 [0] NCCL INFO Connected all trees +ip-26-0-169-207:619094:619378 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619094:619378 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:619094:619378 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619095:619381 [1] NCCL INFO comm 0xad4a5f0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc81bd109e66396f7 - Init COMPLETE +ip-26-0-169-207:619094:619378 [0] NCCL INFO comm 0xbf80f70 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc81bd109e66396f7 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:619095:619398 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:619094:619396 [0] NCCL INFO comm 0xbf94b40 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf93abd0d238b12f9 - Init START +ip-26-0-169-207:619095:619398 [1] NCCL INFO comm 0xad5e220 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x488f0ca8b8a768c3 - Init START +ip-26-0-169-207:619094:619396 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:619095:619398 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619098:619376 [4] NCCL INFO Connected all rings +ip-26-0-169-207:619098:619376 [4] NCCL INFO Connected all trees +ip-26-0-169-207:619098:619376 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619098:619376 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:619098:619376 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619099:619379 [5] NCCL INFO Connected all rings +ip-26-0-169-207:619099:619379 [5] NCCL INFO Connected all trees +ip-26-0-169-207:619099:619379 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619099:619379 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:619099:619379 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619097:619382 [3] NCCL INFO Connected all rings +ip-26-0-169-207:619097:619382 [3] NCCL INFO Connected all trees +ip-26-0-169-207:619097:619382 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619097:619382 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:619097:619382 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619098:619376 [4] NCCL INFO comm 0xc30d5f0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x59df0f41f787b262 - Init COMPLETE +ip-26-0-169-207:619099:619379 [5] NCCL INFO comm 0xc031670 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x59df0f41f787b262 - Init COMPLETE +ip-26-0-169-207:619096:619380 [2] NCCL INFO Connected all rings +ip-26-0-169-207:619096:619380 [2] NCCL INFO Connected all trees +ip-26-0-169-207:619096:619380 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619096:619380 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:619096:619380 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:619099:619404 [5] NCCL INFO comm 0xc044f50 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56d69aa16fda3cb1 - Init START +ip-26-0-169-207:619099:619404 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:619098:619406 [4] NCCL INFO comm 0xc321400 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8a176cdf9fe5e001 - Init START +ip-26-0-169-207:619098:619406 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619097:619382 [3] NCCL INFO comm 0xa5c3cb0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44a96a70fdbb664d - Init COMPLETE +ip-26-0-169-207:619096:619380 [2] NCCL INFO comm 0xad05d30 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44a96a70fdbb664d - Init COMPLETE +ip-26-0-169-207:619100:619375 [6] NCCL INFO Connected all rings +ip-26-0-169-207:619100:619375 [6] NCCL INFO Connected all trees +ip-26-0-169-207:619100:619375 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619100:619375 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:619100:619375 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:619101:619377 [7] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619377 [7] NCCL INFO Connected all trees +ip-26-0-169-207:619101:619377 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619101:619377 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:619101:619377 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619096:619409 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:619097:619411 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:619096:619409 [2] NCCL INFO comm 0xad19c90 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56b75c97c4e37 - Init START +ip-26-0-169-207:619096:619409 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619097:619411 [3] NCCL INFO comm 0xa5d78d0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x84b68d506d29b59d - Init START +ip-26-0-169-207:619097:619411 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619100:619375 [6] NCCL INFO comm 0xa6a6860 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x941a314f46f8864f - Init COMPLETE +ip-26-0-169-207:619101:619377 [7] NCCL INFO comm 0xaa6c220 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x941a314f46f8864f - Init COMPLETE +ip-26-0-169-207:619100:619414 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:619100:619414 [6] NCCL INFO comm 0xa6ba670 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x82db254d0621c66e - Init START +ip-26-0-169-207:619100:619414 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:619101:619416 [7] NCCL INFO comm 0xaa80410 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc54be024cefca611 - Init START +ip-26-0-169-207:619101:619416 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:619094:619396 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:619094:619396 [0] NCCL INFO Connected all rings +ip-26-0-169-207:619094:619396 [0] NCCL INFO Connected all trees +ip-26-0-169-207:619094:619396 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619098:619406 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:619098:619406 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:619098:619406 [4] NCCL INFO Connected all rings +ip-26-0-169-207:619098:619406 [4] NCCL INFO Connected all trees +ip-26-0-169-207:619098:619406 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619095:619398 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:619095:619398 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:619095:619398 [1] NCCL INFO Connected all rings +ip-26-0-169-207:619095:619398 [1] NCCL INFO Connected all trees +ip-26-0-169-207:619095:619398 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619099:619404 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:619099:619404 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:619099:619404 [5] NCCL INFO Connected all rings +ip-26-0-169-207:619099:619404 [5] NCCL INFO Connected all trees +ip-26-0-169-207:619099:619404 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619094:619396 [0] NCCL INFO comm 0xbf94b40 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf93abd0d238b12f9 - Init COMPLETE +ip-26-0-169-207:619098:619406 [4] NCCL INFO comm 0xc321400 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8a176cdf9fe5e001 - Init COMPLETE +ip-26-0-169-207:619095:619398 [1] NCCL INFO comm 0xad5e220 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x488f0ca8b8a768c3 - Init COMPLETE +12/13/2024 10:17:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Total number of parameters: 1.04G (1984.26MiB) +12/13/2024 10:17:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +12/13/2024 10:17:41 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +12/13/2024 10:17:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 15874.00MiB +12/13/2024 10:17:41 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/13/2024 10:17:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: No checkpoint path provided. +12/13/2024 10:17:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Parametrizing model parameters using StandardParametrizator +ip-26-0-169-207:619099:619404 [5] NCCL INFO comm 0xc044f50 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56d69aa16fda3cb1 - Init COMPLETE +ip-26-0-169-207:619094:619427 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:619095:619429 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:619099:619430 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:619098:619431 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:619097:619411 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:619097:619411 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:619097:619411 [3] NCCL INFO Connected all rings +ip-26-0-169-207:619097:619411 [3] NCCL INFO Connected all trees +ip-26-0-169-207:619097:619411 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619096:619409 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:619096:619409 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:619096:619409 [2] NCCL INFO Connected all rings +ip-26-0-169-207:619096:619409 [2] NCCL INFO Connected all trees +ip-26-0-169-207:619096:619409 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619097:619411 [3] NCCL INFO comm 0xa5d78d0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x84b68d506d29b59d - Init COMPLETE +ip-26-0-169-207:619101:619416 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:619101:619416 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:619101:619416 [7] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619416 [7] NCCL INFO Connected all trees +ip-26-0-169-207:619101:619416 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619100:619414 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:619100:619414 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:619100:619414 [6] NCCL INFO Connected all rings +ip-26-0-169-207:619100:619414 [6] NCCL INFO Connected all trees +ip-26-0-169-207:619100:619414 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619096:619409 [2] NCCL INFO comm 0xad19c90 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56b75c97c4e37 - Init COMPLETE +ip-26-0-169-207:619101:619416 [7] NCCL INFO comm 0xaa80410 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc54be024cefca611 - Init COMPLETE +ip-26-0-169-207:619096:619441 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:619097:619442 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:619100:619414 [6] NCCL INFO comm 0xa6ba670 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x82db254d0621c66e - Init COMPLETE +ip-26-0-169-207:619101:619443 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:619101:619443 [7] NCCL INFO comm 0xaa852e0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3dff9e08c31e0a57 - Init START +ip-26-0-169-207:619099:619430 [5] NCCL INFO comm 0xc049ee0 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3dff9e08c31e0a57 - Init START +ip-26-0-169-207:619097:619442 [3] NCCL INFO comm 0xa5dc010 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3dff9e08c31e0a57 - Init START +ip-26-0-169-207:619095:619429 [1] NCCL INFO comm 0xad62ff0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3dff9e08c31e0a57 - Init START +ip-26-0-169-207:619100:619444 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:619101:619443 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619099:619430 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619097:619442 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619095:619429 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619100:619444 [6] NCCL INFO comm 0xa6bfc80 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa31705f17a978d44 - Init START +ip-26-0-169-207:619096:619441 [2] NCCL INFO comm 0xad1ec50 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa31705f17a978d44 - Init START +ip-26-0-169-207:619098:619431 [4] NCCL INFO comm 0xc326c90 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa31705f17a978d44 - Init START +ip-26-0-169-207:619094:619427 [0] NCCL INFO comm 0xbf99cb0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa31705f17a978d44 - Init START +ip-26-0-169-207:619100:619444 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619096:619441 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619098:619431 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:619094:619427 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:619099:619430 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619099:619430 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:619096:619441 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:619096:619441 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:619097:619442 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:619097:619442 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:619098:619431 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619098:619431 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:619100:619444 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619100:619444 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:619101:619443 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:619101:619443 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:619094:619427 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:619095:619429 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:619100:619444 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:619100:619444 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:619094:619427 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619096:619441 [2] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:619096:619441 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619098:619431 [4] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:619098:619431 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:619101:619443 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:619099:619430 [5] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:619101:619443 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:619099:619430 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:619097:619442 [3] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:619097:619442 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:619095:619429 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 04/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 04/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 05/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 04/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 05/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 06/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 05/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 06/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 07/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 06/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 04/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 08/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 07/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 08/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 04/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 07/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 09/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 05/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 05/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 08/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 06/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 09/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 07/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 10/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 09/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 08/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 11/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 10/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 06/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 10/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 09/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 12/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 11/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 07/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 11/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 10/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 13/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 12/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 08/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 12/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 13/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 14/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 13/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 09/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 14/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 15/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 11/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 14/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 16/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 12/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 15/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 17/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 13/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 16/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 18/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 10/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 14/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 17/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 15/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 19/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 11/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 15/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 18/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 16/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 20/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 12/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 16/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 19/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 17/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619095:619429 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 21/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 13/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 17/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 20/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 18/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 22/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 14/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 18/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 21/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 19/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 23/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 15/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 22/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 16/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 19/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 23/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 20/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 17/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 20/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 21/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 18/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 21/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 22/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 19/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 22/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 23/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 23/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 20/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 21/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 22/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 23/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Connected all rings +ip-26-0-169-207:619097:619442 [3] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619443 [7] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Connected all rings +ip-26-0-169-207:619095:619429 [1] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Connected all rings +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Connected all rings +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 02/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 03/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Connected all rings +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 04/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 05/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 06/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 07/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 08/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 09/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 10/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 11/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 12/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 13/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 14/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 15/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619101:619443 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 16/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 17/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 18/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 04/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 19/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 05/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 04/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 20/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 05/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 06/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 21/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 06/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 07/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 22/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 07/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 08/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619100:619444 [6] NCCL INFO Channel 23/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 08/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 09/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 09/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 10/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 10/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 11/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 11/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 12/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 12/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 13/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 13/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 14/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 14/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 15/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 15/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 16/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 16/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 17/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 17/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 18/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619096:619441 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 18/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 19/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 19/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 20/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 20/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 21/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 21/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 22/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 22/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619099:619430 [5] NCCL INFO Channel 23/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:619098:619431 [4] NCCL INFO Channel 23/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:619097:619442 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:619094:619427 [0] NCCL INFO Connected all trees +ip-26-0-169-207:619094:619427 [0] NCCL INFO NVLS comm 0xbf99cb0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:619095:619429 [1] NCCL INFO Connected all trees +ip-26-0-169-207:619095:619429 [1] NCCL INFO NVLS comm 0xad62ff0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:619101:619443 [7] NCCL INFO Connected all trees +ip-26-0-169-207:619101:619443 [7] NCCL INFO NVLS comm 0xaa852e0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:619100:619444 [6] NCCL INFO Connected all trees +ip-26-0-169-207:619100:619444 [6] NCCL INFO NVLS comm 0xa6bfc80 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:619099:619430 [5] NCCL INFO Connected all trees +ip-26-0-169-207:619099:619430 [5] NCCL INFO NVLS comm 0xc049ee0 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:619096:619441 [2] NCCL INFO Connected all trees +ip-26-0-169-207:619096:619441 [2] NCCL INFO NVLS comm 0xad1ec50 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:619097:619442 [3] NCCL INFO Connected all trees +ip-26-0-169-207:619097:619442 [3] NCCL INFO NVLS comm 0xa5dc010 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:619098:619431 [4] NCCL INFO Connected all trees +ip-26-0-169-207:619098:619431 [4] NCCL INFO NVLS comm 0xc326c90 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:619101:619443 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619101:619443 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619101:619443 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:619101:619443 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619097:619442 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619097:619442 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619097:619442 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:619097:619442 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619099:619430 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619099:619430 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619099:619430 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:619099:619430 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619095:619429 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619095:619429 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619095:619429 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:619095:619429 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619101:619443 [7] NCCL INFO comm 0xaa852e0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3dff9e08c31e0a57 - Init COMPLETE +ip-26-0-169-207:619097:619442 [3] NCCL INFO comm 0xa5dc010 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3dff9e08c31e0a57 - Init COMPLETE +ip-26-0-169-207:619099:619430 [5] NCCL INFO comm 0xc049ee0 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3dff9e08c31e0a57 - Init COMPLETE +ip-26-0-169-207:619095:619429 [1] NCCL INFO comm 0xad62ff0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3dff9e08c31e0a57 - Init COMPLETE +ip-26-0-169-207:619100:619444 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619100:619444 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619100:619444 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:619100:619444 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619098:619431 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619098:619431 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619096:619441 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619098:619431 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:619098:619431 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619094:619427 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:619094:619427 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619096:619441 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:619094:619427 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:619096:619441 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:619094:619427 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619096:619441 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:619094:619427 [0] NCCL INFO comm 0xbf99cb0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa31705f17a978d44 - Init COMPLETE +ip-26-0-169-207:619098:619431 [4] NCCL INFO comm 0xc326c90 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa31705f17a978d44 - Init COMPLETE +ip-26-0-169-207:619100:619444 [6] NCCL INFO comm 0xa6bfc80 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa31705f17a978d44 - Init COMPLETE +ip-26-0-169-207:619096:619441 [2] NCCL INFO comm 0xad1ec50 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa31705f17a978d44 - Init COMPLETE +12/13/2024 10:17:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Optimizer Building] Using LearningRateForSP as learning rate +12/13/2024 10:17:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] Size of optimizer params per rank: +12/13/2024 10:17:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 0 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 10:17:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 1 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 10:17:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 2 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 10:17:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 3 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 10:17:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/13/2024 10:17:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Using dummy data generator +12/13/2024 10:17:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] There are 1 training stages +12/13/2024 10:17:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Stage Stable Training Stage] start from step 1 +12/13/2024 10:17:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: +12/13/2024 10:17:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Start training] datetime: 2024-12-13 10:17:46.393042 | mbs: 2 | grad_accum: 256 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/13/2024 10:17:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Waiting for 6 seconds +12/13/2024 10:18:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/13/2024 10:18:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3472.48MiB. Peak allocated 5632.00MiB. Peak reserved: 17860.00MiB +12/13/2024 10:18:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3572.67MiB. Peak allocated 6985.16MiB. Peak reserved: 8202.00MiB +12/13/2024 10:18:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 20K | tokens_per_sec: 210K | tokens_per_sec_per_gpu: 26.2K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 185 | hardware_tflops_per_gpu: 185 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.64G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 10:18:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.85MiB. Peak allocated 6820.14MiB. Peak reserved: 9196.00MiB +12/13/2024 10:18:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.97MiB. Peak allocated 8000.47MiB. Peak reserved: 9452.00MiB +12/13/2024 10:18:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 18.6K | tokens_per_sec: 225K | tokens_per_sec_per_gpu: 28.1K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.0003 | model_tflops_per_gpu: 198 | hardware_tflops_per_gpu: 198 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 11G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 10:18:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.85MiB. Peak allocated 6820.26MiB. Peak reserved: 10446.00MiB +12/13/2024 10:18:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.97MiB. Peak allocated 8000.47MiB. Peak reserved: 10446.00MiB +12/13/2024 10:18:58 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 3 / 100 | consumed_tokens: 12.6M | elapsed_time_per_iteration_ms: 18.7K | tokens_per_sec: 224K | tokens_per_sec_per_gpu: 28K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.000296 | model_tflops_per_gpu: 197 | hardware_tflops_per_gpu: 197 | grad_norm: 0.424 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 11G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 10:18:58 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/13/2024 10:18:58 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: | -------- | -------------------------------------------------------------- | ----- | ------- | --- | ----------- | ---- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | --------------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/13/2024 10:18:58 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: | 13417423 | 1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k | 1 | 2048 | 2 | 256 | 2048 | 197.28 | 197.28 | 27997.53 | 461.37 | 264.40 | 263.52 | 458.60 | 263.68 | 264.47 | 6.66 | 10.20 | 4 | 1 | 2 | | TensorParallelLinearMode.REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 8 | 2048 | 32768 | True | torch.bfloat16 | 1 | 25 | True | 1.04G | 520M | +12/13/2024 10:18:58 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final.csv +12/13/2024 10:18:58 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Acquired lock for CSV file: benchmark/results/bench_final.csv +slurmstepd: error: *** JOB 13417423 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T10:19:03 DUE TO TIME LIMIT *** +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +slurmstepd: error: *** STEP 13417423.0 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T10:19:03 DUE TO TIME LIMIT *** +[2024-12-13 10:19:03,826] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-13 10:19:03,826] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 619094 closing signal SIGTERM +[2024-12-13 10:19:03,826] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 619095 closing signal SIGTERM +[2024-12-13 10:19:03,826] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 619096 closing signal SIGTERM +[2024-12-13 10:19:03,827] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 619097 closing signal SIGTERM +[2024-12-13 10:19:03,829] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 619098 closing signal SIGTERM +[2024-12-13 10:19:03,829] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 619099 closing signal SIGTERM +[2024-12-13 10:19:03,829] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 619100 closing signal SIGTERM +[2024-12-13 10:19:03,829] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 619101 closing signal SIGTERM diff --git a/logs/13417750-bench_stress_test.out b/logs/13417750-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..6a9400fadf9a6c96660a2fbc1c5c09fd59904327 --- /dev/null +++ b/logs/13417750-bench_stress_test.out @@ -0,0 +1,4143 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-160-103 ++ export NODELIST=ip-26-0-160-103 ++ NODELIST=ip-26-0-160-103 +++ scontrol show hostnames ip-26-0-160-103 +++ head -n1 ++ export MASTER_NODE=ip-26-0-160-103 ++ MASTER_NODE=ip-26-0-160-103 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-160-103' +Master node: ip-26-0-160-103 ++ echo 'All nodes: ip-26-0-160-103' +All nodes: ip-26-0-160-103 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13417750 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-103:12356 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 12:26:14,534] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 12:26:14,534] torch.distributed.run: [WARNING] +[2024-12-13 12:26:14,534] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 12:26:14,534] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 12:26:14,534] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-160-103:2564087:2564087 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2564087:2564087 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2564087:2564087 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2564087:2564087 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2564087:2564087 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2564089:2564089 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2564089:2564089 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2564093:2564093 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2564093:2564093 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2564089:2564089 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2564093:2564093 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2564089:2564089 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2564093:2564093 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2564089:2564089 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2564093:2564093 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2564094:2564094 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2564094:2564094 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2564094:2564094 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2564092:2564092 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2564092:2564092 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2564094:2564094 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2564094:2564094 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2564092:2564092 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2564092:2564092 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2564092:2564092 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2564088:2564088 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2564088:2564088 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2564088:2564088 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2564090:2564090 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2564088:2564088 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2564088:2564088 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2564090:2564090 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2564090:2564090 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2564090:2564090 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2564090:2564090 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2564091:2564091 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-160-103:2564091:2564091 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-160-103:2564091:2564091 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.160.103<0> +ip-26-0-160-103:2564091:2564091 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-160-103:2564091:2564091 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.9.1-aws +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Using Libfabric version 1.21 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Running on p5.48xlarge platform, Setting NCCL_TOPO_FILE environment variable to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Using transport protocol RDMA +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO comm 0x84f6120 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb1ba8ba1bed58f26 - Init START +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO comm 0x9e33a90 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb1ba8ba1bed58f26 - Init START +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO comm 0x8cab550 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb1ba8ba1bed58f26 - Init START +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO comm 0x8799fc0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb1ba8ba1bed58f26 - Init START +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO comm 0xa279840 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb1ba8ba1bed58f26 - Init START +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO comm 0x94889c0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb1ba8ba1bed58f26 - Init START +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO comm 0x9291640 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb1ba8ba1bed58f26 - Init START +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO comm 0x9164690 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb1ba8ba1bed58f26 - Init START +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NVLS comm 0x94889c0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NVLS comm 0x8cab550 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NVLS comm 0x84f6120 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NVLS comm 0x9164690 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NVLS comm 0x8799fc0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NVLS comm 0x9e33a90 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NVLS comm 0xa279840 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NVLS comm 0x9291640 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564088:2564215 [1] NCCL INFO comm 0x84f6120 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb1ba8ba1bed58f26 - Init COMPLETE +ip-26-0-160-103:2564092:2564213 [5] NCCL INFO comm 0x9164690 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb1ba8ba1bed58f26 - Init COMPLETE +ip-26-0-160-103:2564090:2564216 [3] NCCL INFO comm 0x8799fc0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb1ba8ba1bed58f26 - Init COMPLETE +ip-26-0-160-103:2564094:2564214 [7] NCCL INFO comm 0x94889c0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb1ba8ba1bed58f26 - Init COMPLETE +ip-26-0-160-103:2564093:2564212 [6] NCCL INFO comm 0x9e33a90 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb1ba8ba1bed58f26 - Init COMPLETE +ip-26-0-160-103:2564091:2564217 [4] NCCL INFO comm 0xa279840 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb1ba8ba1bed58f26 - Init COMPLETE +ip-26-0-160-103:2564087:2564210 [0] NCCL INFO comm 0x8cab550 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb1ba8ba1bed58f26 - Init COMPLETE +ip-26-0-160-103:2564089:2564211 [2] NCCL INFO comm 0x9291640 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb1ba8ba1bed58f26 - Init COMPLETE +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO comm 0x89c3a80 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7da3b3c5d28fc7a - Init START +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO comm 0x8ed5070 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7da3b3c5d28fc7a - Init START +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO comm 0x96b2600 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7da3b3c5d28fc7a - Init START +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO comm 0x94bb240 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7da3b3c5d28fc7a - Init START +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO comm 0x871fa40 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7da3b3c5d28fc7a - Init START +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO comm 0xa05d400 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7da3b3c5d28fc7a - Init START +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO comm 0x938e130 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7da3b3c5d28fc7a - Init START +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO comm 0xa4a3840 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7da3b3c5d28fc7a - Init START +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NVLS comm 0x8ed5070 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NVLS comm 0x96b2600 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NVLS comm 0x89c3a80 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NVLS comm 0x94bb240 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NVLS comm 0xa4a3840 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NVLS comm 0x871fa40 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NVLS comm 0x938e130 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NVLS comm 0xa05d400 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564094:2564285 [7] NCCL INFO comm 0x96b2600 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7da3b3c5d28fc7a - Init COMPLETE +ip-26-0-160-103:2564090:2564287 [3] NCCL INFO comm 0x89c3a80 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7da3b3c5d28fc7a - Init COMPLETE +ip-26-0-160-103:2564092:2564284 [5] NCCL INFO comm 0x938e130 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7da3b3c5d28fc7a - Init COMPLETE +ip-26-0-160-103:2564093:2564283 [6] NCCL INFO comm 0xa05d400 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7da3b3c5d28fc7a - Init COMPLETE +ip-26-0-160-103:2564088:2564282 [1] NCCL INFO comm 0x871fa40 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7da3b3c5d28fc7a - Init COMPLETE +ip-26-0-160-103:2564089:2564286 [2] NCCL INFO comm 0x94bb240 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7da3b3c5d28fc7a - Init COMPLETE +ip-26-0-160-103:2564091:2564281 [4] NCCL INFO comm 0xa4a3840 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7da3b3c5d28fc7a - Init COMPLETE +ip-26-0-160-103:2564087:2564280 [0] NCCL INFO comm 0x8ed5070 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7da3b3c5d28fc7a - Init COMPLETE +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO comm 0x87341c0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa59a128e2017022c - Init START +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO comm 0x94cf990 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa59a128e2017022c - Init START +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO comm 0x96c6c70 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa59a128e2017022c - Init START +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO comm 0x8ee9550 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa59a128e2017022c - Init START +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO comm 0x89d81c0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa59a128e2017022c - Init START +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO comm 0xa071b50 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa59a128e2017022c - Init START +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO comm 0x93a2880 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa59a128e2017022c - Init START +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO comm 0xa4b7f80 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa59a128e2017022c - Init START +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NVLS comm 0x8ee9550 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NVLS comm 0x96c6c70 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NVLS comm 0x87341c0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NVLS comm 0x89d81c0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NVLS comm 0x93a2880 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NVLS comm 0x94cf990 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NVLS comm 0xa4b7f80 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NVLS comm 0xa071b50 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564090:2564317 [3] NCCL INFO comm 0x89d81c0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa59a128e2017022c - Init COMPLETE +ip-26-0-160-103:2564094:2564320 [7] NCCL INFO comm 0x96c6c70 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa59a128e2017022c - Init COMPLETE +ip-26-0-160-103:2564092:2564321 [5] NCCL INFO comm 0x93a2880 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa59a128e2017022c - Init COMPLETE +ip-26-0-160-103:2564088:2564315 [1] NCCL INFO comm 0x87341c0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa59a128e2017022c - Init COMPLETE +ip-26-0-160-103:2564093:2564318 [6] NCCL INFO comm 0xa071b50 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa59a128e2017022c - Init COMPLETE +ip-26-0-160-103:2564091:2564316 [4] NCCL INFO comm 0xa4b7f80 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa59a128e2017022c - Init COMPLETE +ip-26-0-160-103:2564089:2564319 [2] NCCL INFO comm 0x94cf990 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa59a128e2017022c - Init COMPLETE +ip-26-0-160-103:2564087:2564314 [0] NCCL INFO comm 0x8ee9550 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa59a128e2017022c - Init COMPLETE +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Config: +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Config(general=GeneralArgs(project='debug', +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: run='1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k', +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: seed=42, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: step=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: consumed_train_samples=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: ignore_sanity_checks=True), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: parallelism=ParallelismArgs(dp=4, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pp=1, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp=2, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pp_engine=, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_mode=, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_linear_async_communication=True, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: recompute_layer=False, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_recompute_allgather=True, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: expert_parallel_size=1), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: eos_token_id=0, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_act='silu', +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_size=2048, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: initializer_range=0.02, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: intermediate_size=8192, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: is_llama_config=True, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: max_position_embeddings=2048, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_attention_heads=32, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_hidden_layers=16, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_key_value_heads=8, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pad_token_id=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pretraining_tp=1, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rms_norm_eps=1e-05, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_scaling=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_theta=10000.0, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_interleaved=False, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tie_word_embeddings=True, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: use_cache=True, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: vocab_size=32768), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: init_method=RandomInit(std=0.02), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: dtype=torch.bfloat16, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: make_vocab_size_divisible_by=1, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: ddp_bucket_cap_mb=25), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer_revision=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer_max_length=None), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoint_interval=10000, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: save_initial_state=False, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: save_final_state=False, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: resume_checkpoint_path=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoints_path_is_shared_file_system=False), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: logging=LoggingArgs(log_level='info', +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: log_level_replica='info', +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration_step_info_interval=1), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokens=TokensArgs(sequence_length=2048, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: train_steps=100, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: micro_batch_size=2, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: batch_accumulation_per_replica=256, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: val_check_interval=100, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: limit_val_batches=0, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: limit_test_batches=0), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: adam_beta1=0.9, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: adam_beta2=0.95, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: torch_adam_is_fused=True, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: name='adamW'), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: zero_stage=1, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: weight_decay=0.01, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: clip_grad=1.0, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: accumulate_grad_in_fp32=True, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_warmup_steps=2, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_warmup_style='linear', +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_style='cosine', +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_steps=13, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_starting_step=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: min_decay_lr=1e-05)), +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: start_training_step=1, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: data=DataArgs(dataset=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: seed=42, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_loading_workers=1))], +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: profiler=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lighteval=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: s3_upload=None) +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Model Config: +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: LlamaConfig(bos_token_id=0, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: eos_token_id=0, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_act='silu', +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_size=2048, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: initializer_range=0.02, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: intermediate_size=8192, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: is_llama_config=True, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: max_position_embeddings=2048, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_attention_heads=32, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_hidden_layers=16, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_key_value_heads=8, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pad_token_id=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pretraining_tp=1, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rms_norm_eps=1e-05, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_scaling=None, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_theta=10000.0, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_interleaved=False, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tie_word_embeddings=True, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: use_cache=True, +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: vocab_size=32768) +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Building model.. +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Initialize RoPE Theta = 10000.0 +12/13/2024 12:26:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO comm 0xc2bd610 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x866ab0dbbf893e53 - Init START +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO comm 0xa53baa0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x287cbdf49738ae52 - Init START +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO comm 0xb1a7de0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x866ab0dbbf893e53 - Init START +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO comm 0xacf05f0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x287cbdf49738ae52 - Init START +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO comm 0xbe7a3f0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd9877d6dc82ce7a - Init START +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO comm 0xb4ce140 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd9877d6dc82ce7a - Init START +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO comm 0xa7de900 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd63e5aa3e4ce3cdb - Init START +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO comm 0xb2d2a30 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd63e5aa3e4ce3cdb - Init START +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564088:2564374 [1] NCCL INFO comm 0xa53baa0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x287cbdf49738ae52 - Init COMPLETE +ip-26-0-160-103:2564087:2564370 [0] NCCL INFO comm 0xacf05f0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x287cbdf49738ae52 - Init COMPLETE +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO comm 0xa54e760 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x20ea49d68d7a9435 - Init START +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO comm 0xad032b0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1d352f4634dc9585 - Init START +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564094:2564373 [7] NCCL INFO comm 0xb4ce140 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd9877d6dc82ce7a - Init COMPLETE +ip-26-0-160-103:2564093:2564369 [6] NCCL INFO comm 0xbe7a3f0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd9877d6dc82ce7a - Init COMPLETE +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO comm 0xbe8d0b0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc6dbcd71c776b3bb - Init START +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO comm 0xb4e0e00 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb55de174ac92ee9d - Init START +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564091:2564368 [4] NCCL INFO comm 0xc2bd610 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x866ab0dbbf893e53 - Init COMPLETE +ip-26-0-160-103:2564092:2564372 [5] NCCL INFO comm 0xb1a7de0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x866ab0dbbf893e53 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO comm 0xb1baaa0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x58b49f9a29b4f92e - Init START +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO comm 0xc2d02d0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x38a7a9a9ca2da1ea - Init START +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564090:2564376 [3] NCCL INFO comm 0xa7de900 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd63e5aa3e4ce3cdb - Init COMPLETE +ip-26-0-160-103:2564089:2564375 [2] NCCL INFO comm 0xb2d2a30 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd63e5aa3e4ce3cdb - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO comm 0xa7f15c0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x296b038c7f0ad57a - Init START +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO comm 0xb2e6b60 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x59fa90abc0613ef3 - Init START +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564087:2564395 [0] NCCL INFO comm 0xad032b0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1d352f4634dc9585 - Init COMPLETE +ip-26-0-160-103:2564093:2564397 [6] NCCL INFO comm 0xbe8d0b0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc6dbcd71c776b3bb - Init COMPLETE +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564088:2564394 [1] NCCL INFO comm 0xa54e760 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x20ea49d68d7a9435 - Init COMPLETE +12/13/2024 12:26:52 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-103]: Local number of parameters: 520M (992.13MiB) +12/13/2024 12:26:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Total number of parameters: 1.04G (1984.26MiB) +12/13/2024 12:26:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Local number of parameters: 520M (992.13MiB) +12/13/2024 12:26:52 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-103]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/13/2024 12:26:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/13/2024 12:26:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: No checkpoint path provided. +12/13/2024 12:26:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Parametrizing model parameters using StandardParametrizator +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564094:2564399 [7] NCCL INFO comm 0xb4e0e00 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb55de174ac92ee9d - Init COMPLETE +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564091:2564405 [4] NCCL INFO comm 0xc2d02d0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x38a7a9a9ca2da1ea - Init COMPLETE +ip-26-0-160-103:2564092:2564403 [5] NCCL INFO comm 0xb1baaa0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x58b49f9a29b4f92e - Init COMPLETE +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564089:2564411 [2] NCCL INFO comm 0xb2e6b60 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x59fa90abc0613ef3 - Init COMPLETE +ip-26-0-160-103:2564090:2564409 [3] NCCL INFO comm 0xa7f15c0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x296b038c7f0ad57a - Init COMPLETE +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Using network AWS Libfabric +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO comm 0xbe900d0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x679ff834aa5825b8 - Init START +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO comm 0xb2ebb50 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x679ff834aa5825b8 - Init START +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO comm 0xc2d3b10 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x679ff834aa5825b8 - Init START +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO comm 0xad06d40 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x679ff834aa5825b8 - Init START +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO comm 0xb4e4480 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x226267637ae221a4 - Init START +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO comm 0xb1bded0 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x226267637ae221a4 - Init START +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO comm 0xa7f45e0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x226267637ae221a4 - Init START +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO comm 0xa551c30 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x226267637ae221a4 - Init START +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /opt/aws-ofi-nccl/share/aws-ofi-nccl/xml/p5.48xl-topo.xml +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NET/OFI Libfabric provider associates MRs with domains +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 04/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 05/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 04/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 06/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 05/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 07/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 04/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 04/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 04/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 06/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 08/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 05/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 05/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 05/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 07/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 09/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 06/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 06/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 08/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 06/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 10/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 07/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 07/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 07/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 09/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 11/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 08/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 08/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 08/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 10/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 12/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 09/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 09/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 11/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 09/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 13/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 10/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 12/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 11/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 10/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 13/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 12/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 10/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 14/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 11/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 14/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 13/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 11/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 15/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 12/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 15/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 14/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 12/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 16/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 13/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 16/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 15/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 13/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 17/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 14/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 17/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 16/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 14/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 18/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 15/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 18/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 17/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 15/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 19/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 19/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 16/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 18/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 20/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 16/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 20/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 17/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 19/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 17/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 21/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 21/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 18/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 18/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 19/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 20/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 22/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 20/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 23/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 22/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 19/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 23/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 20/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 21/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 21/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 21/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 22/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 22/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 22/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 23/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 23/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 23/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Connected all rings +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Connected all rings +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Connected all rings +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Connected all rings +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Connected all rings +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 02/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 03/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 04/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 05/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 06/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 07/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 08/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 09/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 10/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 11/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 12/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 13/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 14/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 15/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 16/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 17/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 18/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 19/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 20/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 21/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 04/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 22/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 05/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 06/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 07/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 08/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 09/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 10/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 11/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 12/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Channel 23/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 13/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 04/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 14/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 05/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 15/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 06/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 16/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 07/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 17/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 08/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 18/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 19/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 20/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 09/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 21/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 10/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 22/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Channel 23/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 11/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 12/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 13/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 14/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 15/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 16/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 17/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 18/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 19/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 20/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 21/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 22/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Channel 23/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Connected all trees +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NVLS comm 0xa551c30 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Connected all trees +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NVLS comm 0xbe900d0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Connected all trees +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NVLS comm 0xb4e4480 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Connected all trees +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NVLS comm 0xad06d40 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Connected all trees +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NVLS comm 0xb1bded0 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Connected all trees +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NVLS comm 0xa7f45e0 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Connected all trees +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NVLS comm 0xc2d3b10 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Connected all trees +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NVLS comm 0xb2ebb50 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564092:2564430 [5] NCCL INFO comm 0xb1bded0 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x226267637ae221a4 - Init COMPLETE +ip-26-0-160-103:2564088:2564421 [1] NCCL INFO comm 0xa551c30 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x226267637ae221a4 - Init COMPLETE +ip-26-0-160-103:2564094:2564427 [7] NCCL INFO comm 0xb4e4480 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x226267637ae221a4 - Init COMPLETE +ip-26-0-160-103:2564090:2564438 [3] NCCL INFO comm 0xa7f45e0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x226267637ae221a4 - Init COMPLETE +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO Connected NVLS tree +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-160-103:2564093:2564426 [6] NCCL INFO comm 0xbe900d0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x679ff834aa5825b8 - Init COMPLETE +ip-26-0-160-103:2564087:2564423 [0] NCCL INFO comm 0xad06d40 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x679ff834aa5825b8 - Init COMPLETE +ip-26-0-160-103:2564091:2564431 [4] NCCL INFO comm 0xc2d3b10 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x679ff834aa5825b8 - Init COMPLETE +ip-26-0-160-103:2564089:2564437 [2] NCCL INFO comm 0xb2ebb50 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x679ff834aa5825b8 - Init COMPLETE +12/13/2024 12:26:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Optimizer Building] Using LearningRateForSP as learning rate +12/13/2024 12:26:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] Size of optimizer params per rank: +12/13/2024 12:26:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 0 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 12:26:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 1 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 12:26:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 2 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 12:26:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [ZeRO sharding] DP Rank 3 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 12:26:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/13/2024 12:26:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Using dummy data generator +12/13/2024 12:26:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Training Plan] There are 1 training stages +12/13/2024 12:26:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Stage Stable Training Stage] start from step 1 +12/13/2024 12:26:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: +12/13/2024 12:26:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Start training] datetime: 2024-12-13 12:26:57.604304 | mbs: 2 | grad_accum: 256 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/13/2024 12:27:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Waiting for 15 seconds +12/13/2024 12:27:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/13/2024 12:27:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 3472.48MiB. Peak allocated 5632.00MiB. Peak reserved: 18884.00MiB +12/13/2024 12:27:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 3572.67MiB. Peak allocated 6985.16MiB. Peak reserved: 7946.00MiB +12/13/2024 12:27:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 20.9K | tokens_per_sec: 201K | tokens_per_sec_per_gpu: 25.1K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 177 | hardware_tflops_per_gpu: 177 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.64G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.5G | hd_free_memory_tb: 243G +12/13/2024 12:27:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4587.85MiB. Peak allocated 6820.14MiB. Peak reserved: 9190.00MiB +12/13/2024 12:28:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4587.97MiB. Peak allocated 8000.47MiB. Peak reserved: 9190.00MiB +12/13/2024 12:28:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 19.1K | tokens_per_sec: 220K | tokens_per_sec_per_gpu: 27.5K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.0003 | model_tflops_per_gpu: 193 | hardware_tflops_per_gpu: 193 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.64G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.5G | hd_free_memory_tb: 243G +12/13/2024 12:28:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4587.85MiB. Peak allocated 6820.26MiB. Peak reserved: 9190.00MiB +12/13/2024 12:28:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4587.97MiB. Peak allocated 8000.47MiB. Peak reserved: 9190.00MiB +12/13/2024 12:28:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration: 3 / 100 | consumed_tokens: 12.6M | elapsed_time_per_iteration_ms: 19.1K | tokens_per_sec: 220K | tokens_per_sec_per_gpu: 27.4K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.000296 | model_tflops_per_gpu: 193 | hardware_tflops_per_gpu: 193 | grad_norm: 0.424 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.64G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.5G | hd_free_memory_tb: 243G +12/13/2024 12:28:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/13/2024 12:28:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: | -------- | -------------------------------------------------------------- | ----- | ------- | --- | ----------- | ---- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | --------------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/13/2024 12:28:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: | 13417750 | 1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k | 1 | 2048 | 2 | 256 | 2048 | 193.42 | 193.42 | 27449.76 | 459.16 | 265.03 | 264.16 | 459.73 | 264.79 | 264.14 | 6.66 | 8.97 | 4 | 1 | 2 | | TensorParallelLinearMode.REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 8 | 2048 | 32768 | True | torch.bfloat16 | 1 | 25 | True | 1.04G | 520M | +12/13/2024 12:28:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final.csv +12/13/2024 12:28:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Acquired lock for CSV file: benchmark/results/bench_final.csv +slurmstepd: error: *** STEP 13417750.0 ON ip-26-0-160-103 CANCELLED AT 2024-12-13T12:28:29 DUE TO TIME LIMIT *** +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +slurmstepd: error: *** JOB 13417750 ON ip-26-0-160-103 CANCELLED AT 2024-12-13T12:28:29 DUE TO TIME LIMIT *** +[2024-12-13 12:28:29,030] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-13 12:28:29,031] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564087 closing signal SIGTERM +[2024-12-13 12:28:29,031] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564088 closing signal SIGTERM +[2024-12-13 12:28:29,031] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564089 closing signal SIGTERM +[2024-12-13 12:28:29,031] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564090 closing signal SIGTERM +[2024-12-13 12:28:29,031] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564091 closing signal SIGTERM +[2024-12-13 12:28:29,031] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564092 closing signal SIGTERM +[2024-12-13 12:28:29,032] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564093 closing signal SIGTERM +[2024-12-13 12:28:29,032] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564094 closing signal SIGTERM +[2024-12-13 12:28:29,154] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564087 closing signal SIGTERM +[2024-12-13 12:28:29,154] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564088 closing signal SIGTERM +[2024-12-13 12:28:29,154] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564089 closing signal SIGTERM +[2024-12-13 12:28:29,154] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564090 closing signal SIGTERM +[2024-12-13 12:28:29,154] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564091 closing signal SIGTERM +[2024-12-13 12:28:29,154] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564092 closing signal SIGTERM +[2024-12-13 12:28:29,154] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564093 closing signal SIGTERM +[2024-12-13 12:28:29,154] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2564094 closing signal SIGTERM diff --git a/logs/13417860-bench_stress_test.out b/logs/13417860-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..8183d7de099bb1c0ffdebded2e0c88789317fe10 --- /dev/null +++ b/logs/13417860-bench_stress_test.out @@ -0,0 +1,3768 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13417860 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 13:56:21,277] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 13:56:21,277] torch.distributed.run: [WARNING] +[2024-12-13 13:56:21,277] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 13:56:21,277] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 13:56:21,277] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-169-207:714703:714703 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:714703:714703 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:714703:714703 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:714703:714703 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:714703:714703 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:714705:714705 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:714704:714704 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:714708:714708 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:714707:714707 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:714709:714709 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:714704:714704 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:714705:714705 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:714706:714706 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:714708:714708 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:714707:714707 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:714709:714709 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:714706:714706 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:714710:714710 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:714710:714710 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:714704:714704 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:714705:714705 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:714708:714708 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:714707:714707 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:714709:714709 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:714706:714706 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:714710:714710 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:714708:714708 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:714704:714704 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:714707:714707 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:714705:714705 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:714708:714708 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:714704:714704 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:714707:714707 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:714705:714705 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:714706:714706 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:714706:714706 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:714710:714710 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:714710:714710 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:714709:714709 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:714709:714709 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/162 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:714703:714825 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:714703:714825 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:714704:714832 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:714704:714832 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:714707:714827 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:714707:714827 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:714709:714826 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:714709:714826 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:714705:714828 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:714705:714828 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:714708:714831 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:714708:714831 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:714710:714830 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:714710:714830 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:714706:714829 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:714706:714829 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:714706:714829 [3] NCCL INFO comm 0xa2c0000 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x592e11f17ff48b50 - Init START +ip-26-0-169-207:714707:714827 [4] NCCL INFO comm 0xa2d7de0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x592e11f17ff48b50 - Init START +ip-26-0-169-207:714708:714831 [5] NCCL INFO comm 0x898a150 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x592e11f17ff48b50 - Init START +ip-26-0-169-207:714705:714828 [2] NCCL INFO comm 0x84725e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x592e11f17ff48b50 - Init START +ip-26-0-169-207:714710:714830 [7] NCCL INFO comm 0x8b77680 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x592e11f17ff48b50 - Init START +ip-26-0-169-207:714709:714826 [6] NCCL INFO comm 0xa20af70 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x592e11f17ff48b50 - Init START +ip-26-0-169-207:714703:714825 [0] NCCL INFO comm 0xa2c5ad0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x592e11f17ff48b50 - Init START +ip-26-0-169-207:714704:714832 [1] NCCL INFO comm 0x909f330 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x592e11f17ff48b50 - Init START +ip-26-0-169-207:714704:714832 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714705:714828 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714706:714829 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714703:714825 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714707:714827 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714707:714827 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:714703:714825 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:714706:714829 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:714704:714832 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:714706:714829 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:714704:714832 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:714708:714831 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714708:714831 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:714709:714826 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714709:714826 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:714705:714828 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:714705:714828 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:714710:714830 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714710:714830 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:714706:714829 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:714708:714831 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:714706:714829 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714708:714831 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714707:714827 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:714707:714827 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714705:714828 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:714705:714828 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714704:714832 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714704:714832 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714710:714830 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:714703:714825 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714709:714826 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:714709:714826 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714710:714830 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714830 [7] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714825 [0] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Connected all rings +ip-26-0-169-207:714706:714829 [3] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714827 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714826 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714706:714829 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714831 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714828 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714832 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:714830 [7] NCCL INFO Connected all trees +ip-26-0-169-207:714710:714830 [7] NCCL INFO NVLS comm 0x8b77680 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714703:714825 [0] NCCL INFO Connected all trees +ip-26-0-169-207:714703:714825 [0] NCCL INFO NVLS comm 0xa2c5ad0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714709:714826 [6] NCCL INFO Connected all trees +ip-26-0-169-207:714709:714826 [6] NCCL INFO NVLS comm 0xa20af70 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714708:714831 [5] NCCL INFO Connected all trees +ip-26-0-169-207:714708:714831 [5] NCCL INFO NVLS comm 0x898a150 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714704:714832 [1] NCCL INFO Connected all trees +ip-26-0-169-207:714704:714832 [1] NCCL INFO NVLS comm 0x909f330 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714707:714827 [4] NCCL INFO Connected all trees +ip-26-0-169-207:714707:714827 [4] NCCL INFO NVLS comm 0xa2d7de0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714705:714828 [2] NCCL INFO Connected all trees +ip-26-0-169-207:714705:714828 [2] NCCL INFO NVLS comm 0x84725e0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714706:714829 [3] NCCL INFO Connected all trees +ip-26-0-169-207:714706:714829 [3] NCCL INFO NVLS comm 0xa2c0000 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714707:714827 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714709:714826 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714707:714827 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714707:714827 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714707:714827 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714705:714828 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714709:714826 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714709:714826 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714709:714826 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714705:714828 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714705:714828 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714705:714828 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714708:714831 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714708:714831 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714708:714831 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714708:714831 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714704:714832 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714704:714832 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714704:714832 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714704:714832 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714706:714829 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714706:714829 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714706:714829 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714706:714829 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714710:714830 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714710:714830 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714710:714830 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714710:714830 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714703:714825 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714703:714825 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714703:714825 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714703:714825 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714710:714830 [7] NCCL INFO comm 0x8b77680 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x592e11f17ff48b50 - Init COMPLETE +ip-26-0-169-207:714706:714829 [3] NCCL INFO comm 0xa2c0000 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x592e11f17ff48b50 - Init COMPLETE +ip-26-0-169-207:714708:714831 [5] NCCL INFO comm 0x898a150 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x592e11f17ff48b50 - Init COMPLETE +ip-26-0-169-207:714705:714828 [2] NCCL INFO comm 0x84725e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x592e11f17ff48b50 - Init COMPLETE +ip-26-0-169-207:714704:714832 [1] NCCL INFO comm 0x909f330 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x592e11f17ff48b50 - Init COMPLETE +ip-26-0-169-207:714707:714827 [4] NCCL INFO comm 0xa2d7de0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x592e11f17ff48b50 - Init COMPLETE +ip-26-0-169-207:714703:714825 [0] NCCL INFO comm 0xa2c5ad0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x592e11f17ff48b50 - Init COMPLETE +ip-26-0-169-207:714709:714826 [6] NCCL INFO comm 0xa20af70 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x592e11f17ff48b50 - Init COMPLETE +ip-26-0-169-207:714703:714896 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:714709:714897 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:714708:714898 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:714704:714899 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:714705:714900 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:714707:714901 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:714710:714902 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:714706:714903 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:714705:714900 [2] NCCL INFO comm 0x869bed0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb27e501423eb4a6a - Init START +ip-26-0-169-207:714706:714903 [3] NCCL INFO comm 0xa4ea140 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb27e501423eb4a6a - Init START +ip-26-0-169-207:714704:714899 [1] NCCL INFO comm 0x92c8ec0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb27e501423eb4a6a - Init START +ip-26-0-169-207:714710:714902 [7] NCCL INFO comm 0x8da1390 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb27e501423eb4a6a - Init START +ip-26-0-169-207:714707:714901 [4] NCCL INFO comm 0xa501360 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb27e501423eb4a6a - Init START +ip-26-0-169-207:714708:714898 [5] NCCL INFO comm 0x8bb66c0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb27e501423eb4a6a - Init START +ip-26-0-169-207:714709:714897 [6] NCCL INFO comm 0xa436c60 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb27e501423eb4a6a - Init START +ip-26-0-169-207:714703:714896 [0] NCCL INFO comm 0xa4ef6a0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb27e501423eb4a6a - Init START +ip-26-0-169-207:714703:714896 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:714705:714900 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714706:714903 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714709:714897 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714704:714899 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714710:714902 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714708:714898 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714707:714901 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714706:714903 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:714706:714903 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:714705:714900 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:714705:714900 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:714707:714901 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714707:714901 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:714709:714897 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714709:714897 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:714708:714898 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714708:714898 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:714710:714902 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714710:714902 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:714703:714896 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:714704:714899 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:714704:714899 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:714704:714899 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:714704:714899 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714705:714900 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:714705:714900 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714706:714903 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:714706:714903 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714707:714901 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:714707:714901 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714710:714902 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714710:714902 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714708:714898 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714709:714897 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714896 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:714703:714896 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714708:714898 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714709:714897 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Connected all rings +ip-26-0-169-207:714704:714899 [1] NCCL INFO Connected all rings +ip-26-0-169-207:714703:714896 [0] NCCL INFO Connected all rings +ip-26-0-169-207:714708:714898 [5] NCCL INFO Connected all rings +ip-26-0-169-207:714705:714900 [2] NCCL INFO Connected all rings +ip-26-0-169-207:714707:714901 [4] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714902 [7] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714903 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714899 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714900 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714901 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714896 [0] NCCL INFO Connected all trees +ip-26-0-169-207:714703:714896 [0] NCCL INFO NVLS comm 0xa4ef6a0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714897 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714898 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714710:714902 [7] NCCL INFO Connected all trees +ip-26-0-169-207:714710:714902 [7] NCCL INFO NVLS comm 0x8da1390 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714705:714900 [2] NCCL INFO Connected all trees +ip-26-0-169-207:714705:714900 [2] NCCL INFO NVLS comm 0x869bed0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714704:714899 [1] NCCL INFO Connected all trees +ip-26-0-169-207:714704:714899 [1] NCCL INFO NVLS comm 0x92c8ec0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714707:714901 [4] NCCL INFO Connected all trees +ip-26-0-169-207:714707:714901 [4] NCCL INFO NVLS comm 0xa501360 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714706:714903 [3] NCCL INFO Connected all trees +ip-26-0-169-207:714706:714903 [3] NCCL INFO NVLS comm 0xa4ea140 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714709:714897 [6] NCCL INFO Connected all trees +ip-26-0-169-207:714709:714897 [6] NCCL INFO NVLS comm 0xa436c60 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714708:714898 [5] NCCL INFO Connected all trees +ip-26-0-169-207:714708:714898 [5] NCCL INFO NVLS comm 0x8bb66c0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714707:714901 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714707:714901 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714707:714901 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714707:714901 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714703:714896 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714703:714896 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714703:714896 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714703:714896 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714705:714900 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714705:714900 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714705:714900 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714705:714900 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714706:714903 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714706:714903 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714706:714903 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714706:714903 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714708:714898 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714709:714897 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714709:714897 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714708:714898 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714709:714897 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714709:714897 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714708:714898 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714708:714898 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714704:714899 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714704:714899 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714704:714899 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714704:714899 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714710:714902 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714710:714902 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714710:714902 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714710:714902 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714705:714900 [2] NCCL INFO comm 0x869bed0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb27e501423eb4a6a - Init COMPLETE +ip-26-0-169-207:714703:714896 [0] NCCL INFO comm 0xa4ef6a0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb27e501423eb4a6a - Init COMPLETE +ip-26-0-169-207:714707:714901 [4] NCCL INFO comm 0xa501360 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb27e501423eb4a6a - Init COMPLETE +ip-26-0-169-207:714709:714897 [6] NCCL INFO comm 0xa436c60 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb27e501423eb4a6a - Init COMPLETE +ip-26-0-169-207:714706:714903 [3] NCCL INFO comm 0xa4ea140 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb27e501423eb4a6a - Init COMPLETE +ip-26-0-169-207:714710:714902 [7] NCCL INFO comm 0x8da1390 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb27e501423eb4a6a - Init COMPLETE +ip-26-0-169-207:714704:714899 [1] NCCL INFO comm 0x92c8ec0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb27e501423eb4a6a - Init COMPLETE +ip-26-0-169-207:714708:714898 [5] NCCL INFO comm 0x8bb66c0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb27e501423eb4a6a - Init COMPLETE +ip-26-0-169-207:714703:714930 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:714704:714934 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:714710:714933 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:714706:714931 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:714707:714935 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:714708:714936 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:714709:714937 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:714705:714932 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:714707:714935 [4] NCCL INFO comm 0xa515ab0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xbfa29dcd434933b6 - Init START +ip-26-0-169-207:714704:714934 [1] NCCL INFO comm 0x92dd610 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbfa29dcd434933b6 - Init START +ip-26-0-169-207:714706:714931 [3] NCCL INFO comm 0xa4fe860 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbfa29dcd434933b6 - Init START +ip-26-0-169-207:714708:714936 [5] NCCL INFO comm 0x8bc9ba0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbfa29dcd434933b6 - Init START +ip-26-0-169-207:714705:714932 [2] NCCL INFO comm 0x86b0600 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xbfa29dcd434933b6 - Init START +ip-26-0-169-207:714703:714930 [0] NCCL INFO comm 0xa503c10 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xbfa29dcd434933b6 - Init START +ip-26-0-169-207:714709:714937 [6] NCCL INFO comm 0xa44a130 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbfa29dcd434933b6 - Init START +ip-26-0-169-207:714710:714933 [7] NCCL INFO comm 0x8db5a20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbfa29dcd434933b6 - Init START +ip-26-0-169-207:714707:714935 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714704:714934 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714705:714932 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714708:714936 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714706:714931 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714709:714937 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714710:714933 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714703:714930 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:714704:714934 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:714704:714934 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:714703:714930 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:714705:714932 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:714705:714932 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:714708:714936 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714708:714936 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:714706:714931 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:714706:714931 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:714709:714937 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714709:714937 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:714710:714933 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714710:714933 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:714707:714935 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714707:714935 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:714707:714935 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:714707:714935 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714706:714931 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:714708:714936 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:714706:714931 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714704:714934 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:714708:714936 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714709:714937 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:714704:714934 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714705:714932 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:714709:714937 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714710:714933 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:714705:714932 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714710:714933 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:714703:714930 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714703:714930 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Connected all rings +ip-26-0-169-207:714709:714937 [6] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714933 [7] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Connected all rings +ip-26-0-169-207:714704:714934 [1] NCCL INFO Connected all rings +ip-26-0-169-207:714703:714930 [0] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Connected all rings +ip-26-0-169-207:714707:714935 [4] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714704:714934 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714709:714937 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714706:714931 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714707:714935 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714936 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714705:714932 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714933 [7] NCCL INFO Connected all trees +ip-26-0-169-207:714710:714933 [7] NCCL INFO NVLS comm 0x8db5a20 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714703:714930 [0] NCCL INFO Connected all trees +ip-26-0-169-207:714703:714930 [0] NCCL INFO NVLS comm 0xa503c10 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714709:714937 [6] NCCL INFO Connected all trees +ip-26-0-169-207:714709:714937 [6] NCCL INFO NVLS comm 0xa44a130 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714707:714935 [4] NCCL INFO Connected all trees +ip-26-0-169-207:714707:714935 [4] NCCL INFO NVLS comm 0xa515ab0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714708:714936 [5] NCCL INFO Connected all trees +ip-26-0-169-207:714708:714936 [5] NCCL INFO NVLS comm 0x8bc9ba0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714706:714931 [3] NCCL INFO Connected all trees +ip-26-0-169-207:714706:714931 [3] NCCL INFO NVLS comm 0xa4fe860 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714704:714934 [1] NCCL INFO Connected all trees +ip-26-0-169-207:714704:714934 [1] NCCL INFO NVLS comm 0x92dd610 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714705:714932 [2] NCCL INFO Connected all trees +ip-26-0-169-207:714705:714932 [2] NCCL INFO NVLS comm 0x86b0600 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:714708:714936 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714708:714936 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714708:714936 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714708:714936 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714706:714931 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714706:714931 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714706:714931 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714706:714931 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714710:714933 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714710:714933 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714710:714933 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714710:714933 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714707:714935 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714707:714935 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714707:714935 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714707:714935 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714704:714934 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714705:714932 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714705:714932 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714704:714934 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714705:714932 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714705:714932 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714704:714934 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714704:714934 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714709:714937 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714709:714937 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714709:714937 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714709:714937 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714703:714930 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714703:714930 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714703:714930 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:714703:714930 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714710:714933 [7] NCCL INFO comm 0x8db5a20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbfa29dcd434933b6 - Init COMPLETE +ip-26-0-169-207:714706:714931 [3] NCCL INFO comm 0xa4fe860 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbfa29dcd434933b6 - Init COMPLETE +ip-26-0-169-207:714709:714937 [6] NCCL INFO comm 0xa44a130 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbfa29dcd434933b6 - Init COMPLETE +ip-26-0-169-207:714705:714932 [2] NCCL INFO comm 0x86b0600 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xbfa29dcd434933b6 - Init COMPLETE +ip-26-0-169-207:714708:714936 [5] NCCL INFO comm 0x8bc9ba0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbfa29dcd434933b6 - Init COMPLETE +ip-26-0-169-207:714704:714934 [1] NCCL INFO comm 0x92dd610 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbfa29dcd434933b6 - Init COMPLETE +ip-26-0-169-207:714707:714935 [4] NCCL INFO comm 0xa515ab0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xbfa29dcd434933b6 - Init COMPLETE +ip-26-0-169-207:714703:714930 [0] NCCL INFO comm 0xa503c10 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xbfa29dcd434933b6 - Init COMPLETE +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config: +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config(general=GeneralArgs(project='debug', +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: run='1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k', +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: step=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: consumed_train_samples=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ignore_sanity_checks=True), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: parallelism=ParallelismArgs(dp=4, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp=1, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp=2, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp_engine=, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_mode=, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_linear_async_communication=True, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: recompute_layer=False, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_recompute_allgather=True, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: expert_parallel_size=1), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: init_method=RandomInit(std=0.02), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: dtype=torch.bfloat16, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: make_vocab_size_divisible_by=1, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ddp_bucket_cap_mb=25), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_revision=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_max_length=None), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoint_interval=10000, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_initial_state=False, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_final_state=False, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: resume_checkpoint_path=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints_path_is_shared_file_system=False), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: logging=LoggingArgs(log_level='info', +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: log_level_replica='info', +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration_step_info_interval=1), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokens=TokensArgs(sequence_length=2048, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: train_steps=100, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: micro_batch_size=2, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: batch_accumulation_per_replica=256, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: val_check_interval=100, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_val_batches=0, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_test_batches=0), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta1=0.9, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta2=0.95, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: torch_adam_is_fused=True, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: name='adamW'), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: zero_stage=1, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: weight_decay=0.01, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: clip_grad=1.0, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: accumulate_grad_in_fp32=True, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_steps=2, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_style='linear', +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_style='cosine', +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_steps=13, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_starting_step=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: min_decay_lr=1e-05)), +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: start_training_step=1, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data=DataArgs(dataset=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_loading_workers=1))], +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: profiler=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lighteval=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: s3_upload=None) +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Model Config: +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: LlamaConfig(bos_token_id=0, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768) +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Building model.. +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Initialize RoPE Theta = 10000.0 +12/13/2024 13:56:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:714703:714984 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:714705:714987 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:714708:714988 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:714704:714989 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:714710:714990 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:714706:714991 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:714709:714986 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:714707:714985 [4] NCCL INFO comm 0xc3172f0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe47703882d67371 - Init START +ip-26-0-169-207:714708:714988 [5] NCCL INFO comm 0xa9cb330 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe47703882d67371 - Init START +ip-26-0-169-207:714703:714984 [0] NCCL INFO comm 0xc305d60 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xea1bb74286bda811 - Init START +ip-26-0-169-207:714704:714989 [1] NCCL INFO comm 0xb0dfa70 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xea1bb74286bda811 - Init START +ip-26-0-169-207:714707:714985 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714705:714987 [2] NCCL INFO comm 0xa4b42b0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x35b33eea34208218 - Init START +ip-26-0-169-207:714706:714991 [3] NCCL INFO comm 0xc2fec70 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x35b33eea34208218 - Init START +ip-26-0-169-207:714703:714984 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:714705:714987 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714710:714990 [7] NCCL INFO comm 0xabb7330 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1caa19182d8b364a - Init START +ip-26-0-169-207:714709:714986 [6] NCCL INFO comm 0xc249970 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1caa19182d8b364a - Init START +ip-26-0-169-207:714709:714986 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714708:714988 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714704:714989 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714706:714991 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714710:714990 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:714707:714985 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714706:714991 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:714706:714991 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:714706:714991 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:714705:714987 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:714708:714988 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:714708:714988 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:714708:714988 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:714707:714985 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714704:714989 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:714704:714989 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:714704:714989 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714710:714990 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714710:714990 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:714710:714990 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:714709:714986 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:714703:714984 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:714987 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714708:714988 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:714985 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714704:714989 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:714986 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:714710:714990 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714703:714984 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:714706:714991 [3] NCCL INFO Connected all rings +ip-26-0-169-207:714706:714991 [3] NCCL INFO Connected all trees +ip-26-0-169-207:714706:714991 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714706:714991 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:714706:714991 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714705:714987 [2] NCCL INFO Connected all rings +ip-26-0-169-207:714705:714987 [2] NCCL INFO Connected all trees +ip-26-0-169-207:714705:714987 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714705:714987 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:714705:714987 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714707:714985 [4] NCCL INFO Connected all rings +ip-26-0-169-207:714707:714985 [4] NCCL INFO Connected all trees +ip-26-0-169-207:714707:714985 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714707:714985 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:714707:714985 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714706:714991 [3] NCCL INFO comm 0xc2fec70 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x35b33eea34208218 - Init COMPLETE +ip-26-0-169-207:714705:714987 [2] NCCL INFO comm 0xa4b42b0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x35b33eea34208218 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:714706:715009 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:714705:715008 [2] NCCL INFO comm 0xa4c83e0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xab2108011f8a531b - Init START +ip-26-0-169-207:714706:715009 [3] NCCL INFO comm 0xc3128a0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf9a81b6eec9070e4 - Init START +ip-26-0-169-207:714705:715008 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714706:715009 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714708:714988 [5] NCCL INFO Connected all rings +ip-26-0-169-207:714708:714988 [5] NCCL INFO Connected all trees +ip-26-0-169-207:714708:714988 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714708:714988 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:714708:714988 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714708:714988 [5] NCCL INFO comm 0xa9cb330 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe47703882d67371 - Init COMPLETE +ip-26-0-169-207:714707:714985 [4] NCCL INFO comm 0xc3172f0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe47703882d67371 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:714708:715014 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:714707:715012 [4] NCCL INFO comm 0xc32b0d0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x63dd9347ed34d033 - Init START +ip-26-0-169-207:714707:715012 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714708:715014 [5] NCCL INFO comm 0xa9def60 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeff69c7b03774bb4 - Init START +ip-26-0-169-207:714708:715014 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714710:714990 [7] NCCL INFO Connected all rings +ip-26-0-169-207:714710:714990 [7] NCCL INFO Connected all trees +ip-26-0-169-207:714710:714990 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714710:714990 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:714710:714990 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714709:714986 [6] NCCL INFO Connected all rings +ip-26-0-169-207:714709:714986 [6] NCCL INFO Connected all trees +ip-26-0-169-207:714709:714986 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714709:714986 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:714709:714986 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714710:714990 [7] NCCL INFO comm 0xabb7330 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1caa19182d8b364a - Init COMPLETE +ip-26-0-169-207:714709:714986 [6] NCCL INFO comm 0xc249970 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1caa19182d8b364a - Init COMPLETE +ip-26-0-169-207:714703:714984 [0] NCCL INFO Connected all rings +ip-26-0-169-207:714703:714984 [0] NCCL INFO Connected all trees +ip-26-0-169-207:714703:714984 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714703:714984 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:714703:714984 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:714710:715019 [7] NCCL INFO comm 0xabcb320 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcfb2f3dc1159ba12 - Init START +ip-26-0-169-207:714710:715019 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:714709:715021 [6] NCCL INFO comm 0xc25da90 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x75f8990b710cd3 - Init START +ip-26-0-169-207:714709:715021 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714704:714989 [1] NCCL INFO Connected all rings +ip-26-0-169-207:714704:714989 [1] NCCL INFO Connected all trees +ip-26-0-169-207:714704:714989 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714704:714989 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:714704:714989 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714704:714989 [1] NCCL INFO comm 0xb0dfa70 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xea1bb74286bda811 - Init COMPLETE +ip-26-0-169-207:714703:714984 [0] NCCL INFO comm 0xc305d60 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xea1bb74286bda811 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:714704:715024 [1] NCCL INFO comm 0xb0f36a0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd7f6c2e6ca4021be - Init START +ip-26-0-169-207:714704:715024 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:714703:715026 [0] NCCL INFO comm 0xc319be0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x140223ff2e40bec7 - Init START +ip-26-0-169-207:714703:715026 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:714705:715008 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:714705:715008 [2] NCCL INFO Connected all rings +ip-26-0-169-207:714705:715008 [2] NCCL INFO Connected all trees +ip-26-0-169-207:714705:715008 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714705:715008 [2] NCCL INFO comm 0xa4c83e0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xab2108011f8a531b - Init COMPLETE +ip-26-0-169-207:714706:715009 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:714706:715009 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:714706:715009 [3] NCCL INFO Connected all rings +ip-26-0-169-207:714706:715009 [3] NCCL INFO Connected all trees +ip-26-0-169-207:714706:715009 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714707:715012 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:714707:715012 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:714707:715012 [4] NCCL INFO Connected all rings +ip-26-0-169-207:714707:715012 [4] NCCL INFO Connected all trees +ip-26-0-169-207:714707:715012 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714706:715009 [3] NCCL INFO comm 0xc3128a0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf9a81b6eec9070e4 - Init COMPLETE +ip-26-0-169-207:714707:715012 [4] NCCL INFO comm 0xc32b0d0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x63dd9347ed34d033 - Init COMPLETE +ip-26-0-169-207:714708:715014 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:714708:715014 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:714708:715014 [5] NCCL INFO Connected all rings +ip-26-0-169-207:714708:715014 [5] NCCL INFO Connected all trees +ip-26-0-169-207:714708:715014 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714708:715014 [5] NCCL INFO comm 0xa9def60 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeff69c7b03774bb4 - Init COMPLETE +ip-26-0-169-207:714710:715019 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:714710:715019 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:714710:715019 [7] NCCL INFO Connected all rings +ip-26-0-169-207:714710:715019 [7] NCCL INFO Connected all trees +ip-26-0-169-207:714710:715019 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714710:715019 [7] NCCL INFO comm 0xabcb320 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcfb2f3dc1159ba12 - Init COMPLETE +ip-26-0-169-207:714703:715026 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:714703:715026 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:714703:715026 [0] NCCL INFO Connected all rings +ip-26-0-169-207:714703:715026 [0] NCCL INFO Connected all trees +ip-26-0-169-207:714703:715026 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714709:715021 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:714709:715021 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:714709:715021 [6] NCCL INFO Connected all rings +ip-26-0-169-207:714709:715021 [6] NCCL INFO Connected all trees +ip-26-0-169-207:714709:715021 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714704:715024 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:714704:715024 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:714704:715024 [1] NCCL INFO Connected all rings +ip-26-0-169-207:714704:715024 [1] NCCL INFO Connected all trees +ip-26-0-169-207:714704:715024 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714703:715026 [0] NCCL INFO comm 0xc319be0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x140223ff2e40bec7 - Init COMPLETE +ip-26-0-169-207:714709:715021 [6] NCCL INFO comm 0xc25da90 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x75f8990b710cd3 - Init COMPLETE +ip-26-0-169-207:714704:715024 [1] NCCL INFO comm 0xb0f36a0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd7f6c2e6ca4021be - Init COMPLETE +12/13/2024 13:56:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Total number of parameters: 1.04G (1984.26MiB) +12/13/2024 13:56:59 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +12/13/2024 13:56:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +12/13/2024 13:56:59 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/13/2024 13:56:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 15874.00MiB +12/13/2024 13:56:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: No checkpoint path provided. +12/13/2024 13:56:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Parametrizing model parameters using StandardParametrizator +ip-26-0-169-207:714708:715047 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:714704:715044 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:714706:715045 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:714703:715049 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:714707:715050 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:714710:715048 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:714705:715052 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:714709:715051 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:714706:715045 [3] NCCL INFO comm 0xc317ac0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb17a043efe8e7fdf - Init START +ip-26-0-169-207:714710:715048 [7] NCCL INFO comm 0xabcfaa0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb17a043efe8e7fdf - Init START +ip-26-0-169-207:714708:715047 [5] NCCL INFO comm 0xa9e40d0 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb17a043efe8e7fdf - Init START +ip-26-0-169-207:714704:715044 [1] NCCL INFO comm 0xb0f87a0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb17a043efe8e7fdf - Init START +ip-26-0-169-207:714710:715048 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714706:715045 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714708:715047 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714704:715044 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714709:715051 [6] NCCL INFO comm 0xc262ac0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8f86b49dccf98b36 - Init START +ip-26-0-169-207:714707:715050 [4] NCCL INFO comm 0xc330780 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8f86b49dccf98b36 - Init START +ip-26-0-169-207:714705:715052 [2] NCCL INFO comm 0xa4cd050 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8f86b49dccf98b36 - Init START +ip-26-0-169-207:714703:715049 [0] NCCL INFO comm 0xc31f190 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8f86b49dccf98b36 - Init START +ip-26-0-169-207:714707:715050 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714705:715052 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714709:715051 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:714703:715049 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:714710:715048 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714710:715048 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:714707:715050 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714707:715050 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:714709:715051 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714709:715051 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:714705:715052 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:714705:715052 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:714703:715049 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:714708:715047 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:714708:715047 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:714704:715044 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:714706:715045 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:714707:715050 [4] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:714707:715050 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714705:715052 [2] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:714709:715051 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:714705:715052 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:714709:715051 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:714703:715049 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714706:715045 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:714708:715047 [5] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:714708:715047 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:714706:715045 [3] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:714706:715045 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:714710:715048 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:714710:715048 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:714704:715044 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 04/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 04/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 04/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 05/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 05/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 06/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 06/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 05/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 07/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 06/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 07/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 08/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 07/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 08/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 08/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 09/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 09/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 09/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 10/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 10/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 10/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 11/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 11/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 12/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 13/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 14/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 11/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 15/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 12/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 12/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 13/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 13/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 14/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 14/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 15/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 15/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 16/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 16/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 16/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 17/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 17/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 18/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 17/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 18/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 19/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 18/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 19/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 20/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 19/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 20/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 21/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 20/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 21/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 22/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 21/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 22/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 23/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 22/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 23/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 23/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 04/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 05/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 06/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 07/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 08/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 09/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 10/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 11/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 12/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 13/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 04/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 14/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 05/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 15/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 16/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 06/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 17/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 18/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 07/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 19/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 08/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 09/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 10/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 20/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 11/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 21/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 12/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 22/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 13/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 23/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 14/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 15/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 16/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 17/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 18/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Connected all rings +ip-26-0-169-207:714704:715044 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 19/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 20/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Connected all rings +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 21/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 22/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 23/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Connected all rings +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Connected all rings +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 02/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 03/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 04/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 05/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 06/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 07/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 08/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 09/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 10/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 11/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 12/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 13/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 14/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 15/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 16/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 17/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 18/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 19/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 20/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Connected all rings +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 21/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714704:715044 [1] NCCL INFO Connected all rings +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 04/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Connected all rings +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 22/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Connected all rings +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 05/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Channel 23/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 06/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 07/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 08/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 09/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 10/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 11/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 12/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 13/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 14/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 15/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 16/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 17/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 18/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 19/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 20/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 21/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 22/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Channel 23/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714710:715048 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 04/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 05/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 06/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 07/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 08/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 09/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 10/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 11/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 12/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714703:715049 [0] NCCL INFO Connected all trees +ip-26-0-169-207:714703:715049 [0] NCCL INFO NVLS comm 0xc31f190 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 13/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714709:715051 [6] NCCL INFO Connected all trees +ip-26-0-169-207:714709:715051 [6] NCCL INFO NVLS comm 0xc262ac0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 14/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 15/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 16/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 17/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 18/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 19/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 20/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 21/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 22/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714708:715047 [5] NCCL INFO Channel 23/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714707:715050 [4] NCCL INFO Connected all trees +ip-26-0-169-207:714707:715050 [4] NCCL INFO NVLS comm 0xc330780 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714706:715045 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:714705:715052 [2] NCCL INFO Connected all trees +ip-26-0-169-207:714705:715052 [2] NCCL INFO NVLS comm 0xa4cd050 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:714703:715049 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714703:715049 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714703:715049 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:714703:715049 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714709:715051 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714709:715051 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714709:715051 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:714709:715051 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714707:715050 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714707:715050 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714707:715050 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:714707:715050 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714705:715052 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714705:715052 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714705:715052 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:714705:715052 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714703:715049 [0] NCCL INFO comm 0xc31f190 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8f86b49dccf98b36 - Init COMPLETE +ip-26-0-169-207:714707:715050 [4] NCCL INFO comm 0xc330780 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8f86b49dccf98b36 - Init COMPLETE +ip-26-0-169-207:714709:715051 [6] NCCL INFO comm 0xc262ac0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8f86b49dccf98b36 - Init COMPLETE +ip-26-0-169-207:714705:715052 [2] NCCL INFO comm 0xa4cd050 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8f86b49dccf98b36 - Init COMPLETE +ip-26-0-169-207:714710:715048 [7] NCCL INFO Connected all trees +ip-26-0-169-207:714710:715048 [7] NCCL INFO NVLS comm 0xabcfaa0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:714704:715044 [1] NCCL INFO Connected all trees +ip-26-0-169-207:714704:715044 [1] NCCL INFO NVLS comm 0xb0f87a0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:714706:715045 [3] NCCL INFO Connected all trees +ip-26-0-169-207:714706:715045 [3] NCCL INFO NVLS comm 0xc317ac0 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:714708:715047 [5] NCCL INFO Connected all trees +ip-26-0-169-207:714708:715047 [5] NCCL INFO NVLS comm 0xa9e40d0 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:714708:715047 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714708:715047 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714708:715047 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:714708:715047 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714710:715048 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714710:715048 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714710:715048 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:714710:715048 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714704:715044 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714704:715044 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714704:715044 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:714704:715044 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714706:715045 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:714706:715045 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:714706:715045 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:714706:715045 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:714704:715044 [1] NCCL INFO comm 0xb0f87a0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb17a043efe8e7fdf - Init COMPLETE +ip-26-0-169-207:714710:715048 [7] NCCL INFO comm 0xabcfaa0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb17a043efe8e7fdf - Init COMPLETE +ip-26-0-169-207:714708:715047 [5] NCCL INFO comm 0xa9e40d0 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb17a043efe8e7fdf - Init COMPLETE +ip-26-0-169-207:714706:715045 [3] NCCL INFO comm 0xc317ac0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb17a043efe8e7fdf - Init COMPLETE +12/13/2024 13:57:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Optimizer Building] Using LearningRateForSP as learning rate +12/13/2024 13:57:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] Size of optimizer params per rank: +12/13/2024 13:57:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 0 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 13:57:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 1 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 13:57:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 2 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 13:57:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 3 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 13:57:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/13/2024 13:57:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Using dummy data generator +12/13/2024 13:57:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] There are 1 training stages +12/13/2024 13:57:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Stage Stable Training Stage] start from step 1 +12/13/2024 13:57:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: +12/13/2024 13:57:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Start training] datetime: 2024-12-13 13:57:04.191492 | mbs: 2 | grad_accum: 256 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/13/2024 13:57:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Waiting for 17 seconds +12/13/2024 13:57:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/13/2024 13:57:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3472.48MiB. Peak allocated 5632.00MiB. Peak reserved: 17860.00MiB +12/13/2024 13:57:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3572.67MiB. Peak allocated 6985.16MiB. Peak reserved: 8202.00MiB +12/13/2024 13:57:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 20K | tokens_per_sec: 210K | tokens_per_sec_per_gpu: 26.2K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 185 | hardware_tflops_per_gpu: 185 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.64G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 13:57:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.85MiB. Peak allocated 6820.14MiB. Peak reserved: 9196.00MiB +12/13/2024 13:58:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.97MiB. Peak allocated 8000.47MiB. Peak reserved: 9452.00MiB +12/13/2024 13:58:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 18.4K | tokens_per_sec: 228K | tokens_per_sec_per_gpu: 28.6K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.0003 | model_tflops_per_gpu: 201 | hardware_tflops_per_gpu: 201 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 11G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 13:58:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.85MiB. Peak allocated 6820.26MiB. Peak reserved: 10446.00MiB +slurmstepd: error: *** JOB 13417860 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T13:58:16 DUE TO TIME LIMIT *** +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +slurmstepd: error: *** STEP 13417860.0 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T13:58:16 DUE TO TIME LIMIT *** +[2024-12-13 13:58:16,291] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-13 13:58:16,291] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714703 closing signal SIGTERM +[2024-12-13 13:58:16,291] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714704 closing signal SIGTERM +[2024-12-13 13:58:16,291] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714705 closing signal SIGTERM +[2024-12-13 13:58:16,291] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714706 closing signal SIGTERM +[2024-12-13 13:58:16,292] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714707 closing signal SIGTERM +[2024-12-13 13:58:16,293] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714708 closing signal SIGTERM +[2024-12-13 13:58:16,295] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714709 closing signal SIGTERM +[2024-12-13 13:58:16,295] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714710 closing signal SIGTERM +[2024-12-13 13:58:16,451] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714703 closing signal SIGTERM +[2024-12-13 13:58:16,451] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714704 closing signal SIGTERM +[2024-12-13 13:58:16,451] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714705 closing signal SIGTERM +[2024-12-13 13:58:16,451] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714706 closing signal SIGTERM +[2024-12-13 13:58:16,451] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714707 closing signal SIGTERM +[2024-12-13 13:58:16,451] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714708 closing signal SIGTERM +[2024-12-13 13:58:16,451] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714709 closing signal SIGTERM +[2024-12-13 13:58:16,451] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 714710 closing signal SIGTERM diff --git a/logs/13417921-bench_stress_test.out b/logs/13417921-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..37215129587832b2f3b77b50d4a1eb2c3619bdf7 --- /dev/null +++ b/logs/13417921-bench_stress_test.out @@ -0,0 +1,3775 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13417921 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 15:10:10,676] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 15:10:10,676] torch.distributed.run: [WARNING] +[2024-12-13 15:10:10,676] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 15:10:10,676] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 15:10:10,676] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-169-207:745720:745720 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:745720:745720 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:745720:745720 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:745720:745720 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:745720:745720 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:745723:745723 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:745721:745721 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:745723:745723 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:745721:745721 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:745723:745723 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:745721:745721 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:745721:745721 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:745723:745723 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:745723:745723 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:745721:745721 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:745724:745724 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:745724:745724 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:745726:745726 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:745726:745726 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:745724:745724 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:745726:745726 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:745724:745724 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:745724:745724 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:745722:745722 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:745725:745725 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:745726:745726 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:745726:745726 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:745722:745722 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:745727:745727 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:745725:745725 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:745727:745727 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:745722:745722 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:745725:745725 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:745727:745727 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:745722:745722 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:745725:745725 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:745722:745722 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:745725:745725 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:745727:745727 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:745727:745727 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/162 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:745720:745844 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:745720:745844 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:745722:745847 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:745722:745847 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:745726:745848 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:745726:745848 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:745721:745842 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:745721:745842 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:745724:745849 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:745724:745849 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:745725:745845 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:745725:745845 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:745723:745843 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:745723:745843 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:745727:745846 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:745727:745846 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:745727:745846 [7] NCCL INFO comm 0xa39ec50 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x618382021d248ea9 - Init START +ip-26-0-169-207:745725:745845 [5] NCCL INFO comm 0x8471ab0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x618382021d248ea9 - Init START +ip-26-0-169-207:745724:745849 [4] NCCL INFO comm 0xa388720 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x618382021d248ea9 - Init START +ip-26-0-169-207:745721:745842 [1] NCCL INFO comm 0x8789a90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x618382021d248ea9 - Init START +ip-26-0-169-207:745723:745843 [3] NCCL INFO comm 0x902aa90 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x618382021d248ea9 - Init START +ip-26-0-169-207:745726:745848 [6] NCCL INFO comm 0x9a642d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x618382021d248ea9 - Init START +ip-26-0-169-207:745720:745844 [0] NCCL INFO comm 0x90eb780 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x618382021d248ea9 - Init START +ip-26-0-169-207:745722:745847 [2] NCCL INFO comm 0x989ad30 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x618382021d248ea9 - Init START +ip-26-0-169-207:745724:745849 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745721:745842 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745723:745843 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745722:745847 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745720:745844 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:745720:745844 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:745724:745849 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745724:745849 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:745722:745847 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:745722:745847 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:745726:745848 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745726:745848 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:745725:745845 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745725:745845 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:745727:745846 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745727:745846 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:745723:745843 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:745723:745843 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:745721:745842 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:745721:745842 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:745722:745847 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:745721:745842 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:745722:745847 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745721:745842 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745726:745848 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745726:745848 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745723:745843 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745724:745849 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:745723:745843 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745724:745849 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745725:745845 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745725:745845 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745727:745846 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745727:745846 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:745720:745844 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745844 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Connected all rings +ip-26-0-169-207:745723:745843 [3] NCCL INFO Connected all rings +ip-26-0-169-207:745725:745845 [5] NCCL INFO Connected all rings +ip-26-0-169-207:745720:745844 [0] NCCL INFO Connected all rings +ip-26-0-169-207:745726:745848 [6] NCCL INFO Connected all rings +ip-26-0-169-207:745727:745846 [7] NCCL INFO Connected all rings +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Connected all rings +ip-26-0-169-207:745721:745842 [1] NCCL INFO Connected all rings +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745849 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745845 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745843 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745848 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745847 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745842 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745846 [7] NCCL INFO Connected all trees +ip-26-0-169-207:745727:745846 [7] NCCL INFO NVLS comm 0xa39ec50 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745726:745848 [6] NCCL INFO Connected all trees +ip-26-0-169-207:745726:745848 [6] NCCL INFO NVLS comm 0x9a642d0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745720:745844 [0] NCCL INFO Connected all trees +ip-26-0-169-207:745720:745844 [0] NCCL INFO NVLS comm 0x90eb780 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745724:745849 [4] NCCL INFO Connected all trees +ip-26-0-169-207:745724:745849 [4] NCCL INFO NVLS comm 0xa388720 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745725:745845 [5] NCCL INFO Connected all trees +ip-26-0-169-207:745725:745845 [5] NCCL INFO NVLS comm 0x8471ab0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745723:745843 [3] NCCL INFO Connected all trees +ip-26-0-169-207:745723:745843 [3] NCCL INFO NVLS comm 0x902aa90 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745722:745847 [2] NCCL INFO Connected all trees +ip-26-0-169-207:745722:745847 [2] NCCL INFO NVLS comm 0x989ad30 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745721:745842 [1] NCCL INFO Connected all trees +ip-26-0-169-207:745721:745842 [1] NCCL INFO NVLS comm 0x8789a90 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745727:745846 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745727:745846 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745727:745846 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745727:745846 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745725:745845 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745723:745843 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745725:745845 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745725:745845 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745725:745845 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745721:745842 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745723:745843 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745721:745842 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745723:745843 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745723:745843 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745721:745842 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745721:745842 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745724:745849 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745724:745849 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745720:745844 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745724:745849 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745724:745849 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745720:745844 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745720:745844 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745720:745844 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745722:745847 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745722:745847 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745722:745847 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745722:745847 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745726:745848 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745726:745848 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745726:745848 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745726:745848 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745727:745846 [7] NCCL INFO comm 0xa39ec50 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x618382021d248ea9 - Init COMPLETE +ip-26-0-169-207:745725:745845 [5] NCCL INFO comm 0x8471ab0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x618382021d248ea9 - Init COMPLETE +ip-26-0-169-207:745726:745848 [6] NCCL INFO comm 0x9a642d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x618382021d248ea9 - Init COMPLETE +ip-26-0-169-207:745723:745843 [3] NCCL INFO comm 0x902aa90 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x618382021d248ea9 - Init COMPLETE +ip-26-0-169-207:745721:745842 [1] NCCL INFO comm 0x8789a90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x618382021d248ea9 - Init COMPLETE +ip-26-0-169-207:745722:745847 [2] NCCL INFO comm 0x989ad30 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x618382021d248ea9 - Init COMPLETE +ip-26-0-169-207:745724:745849 [4] NCCL INFO comm 0xa388720 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x618382021d248ea9 - Init COMPLETE +ip-26-0-169-207:745720:745844 [0] NCCL INFO comm 0x90eb780 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x618382021d248ea9 - Init COMPLETE +ip-26-0-169-207:745720:745913 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:745722:745914 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:745725:745915 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:745723:745916 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:745727:745917 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:745724:745918 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:745721:745919 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:745726:745920 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:745727:745917 [7] NCCL INFO comm 0xa5c85c0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8f20313144f70574 - Init START +ip-26-0-169-207:745724:745918 [4] NCCL INFO comm 0xa5b25e0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8f20313144f70574 - Init START +ip-26-0-169-207:745726:745920 [6] NCCL INFO comm 0x9c8e160 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8f20313144f70574 - Init START +ip-26-0-169-207:745721:745919 [1] NCCL INFO comm 0x89b3ac0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8f20313144f70574 - Init START +ip-26-0-169-207:745723:745916 [3] NCCL INFO comm 0x9254c00 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8f20313144f70574 - Init START +ip-26-0-169-207:745725:745915 [5] NCCL INFO comm 0x869bf70 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8f20313144f70574 - Init START +ip-26-0-169-207:745720:745913 [0] NCCL INFO comm 0x9315770 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8f20313144f70574 - Init START +ip-26-0-169-207:745722:745914 [2] NCCL INFO comm 0x9ac4ec0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8f20313144f70574 - Init START +ip-26-0-169-207:745727:745917 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745724:745918 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745725:745915 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745723:745916 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745721:745919 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745720:745913 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:745726:745920 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745722:745914 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745721:745919 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:745721:745919 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:745724:745918 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745724:745918 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:745727:745917 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745727:745917 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:745726:745920 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745726:745920 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:745722:745914 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:745722:745914 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:745725:745915 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745725:745915 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:745720:745913 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:745723:745916 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:745723:745916 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745723:745916 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745722:745914 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:745723:745916 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745721:745919 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:745722:745914 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745721:745919 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745727:745917 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745724:745918 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745724:745918 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745727:745917 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745726:745920 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745726:745920 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:745720:745913 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745725:745915 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:745725:745915 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Connected all rings +ip-26-0-169-207:745722:745914 [2] NCCL INFO Connected all rings +ip-26-0-169-207:745721:745919 [1] NCCL INFO Connected all rings +ip-26-0-169-207:745727:745917 [7] NCCL INFO Connected all rings +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Connected all rings +ip-26-0-169-207:745724:745918 [4] NCCL INFO Connected all rings +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Connected all rings +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Connected all rings +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745917 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745914 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745918 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745723:745916 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745919 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745920 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745725:745915 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745913 [0] NCCL INFO Connected all trees +ip-26-0-169-207:745720:745913 [0] NCCL INFO NVLS comm 0x9315770 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745727:745917 [7] NCCL INFO Connected all trees +ip-26-0-169-207:745727:745917 [7] NCCL INFO NVLS comm 0xa5c85c0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745722:745914 [2] NCCL INFO Connected all trees +ip-26-0-169-207:745722:745914 [2] NCCL INFO NVLS comm 0x9ac4ec0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745721:745919 [1] NCCL INFO Connected all trees +ip-26-0-169-207:745721:745919 [1] NCCL INFO NVLS comm 0x89b3ac0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745726:745920 [6] NCCL INFO Connected all trees +ip-26-0-169-207:745726:745920 [6] NCCL INFO NVLS comm 0x9c8e160 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745723:745916 [3] NCCL INFO Connected all trees +ip-26-0-169-207:745723:745916 [3] NCCL INFO NVLS comm 0x9254c00 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745724:745918 [4] NCCL INFO Connected all trees +ip-26-0-169-207:745724:745918 [4] NCCL INFO NVLS comm 0xa5b25e0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745725:745915 [5] NCCL INFO Connected all trees +ip-26-0-169-207:745725:745915 [5] NCCL INFO NVLS comm 0x869bf70 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745720:745913 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745720:745913 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745720:745913 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745720:745913 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745722:745914 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745722:745914 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745722:745914 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745722:745914 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745726:745920 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745726:745920 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745726:745920 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745726:745920 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745727:745917 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745727:745917 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745727:745917 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745727:745917 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745724:745918 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745724:745918 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745724:745918 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745724:745918 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745723:745916 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745723:745916 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745723:745916 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745723:745916 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745725:745915 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745725:745915 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745725:745915 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745725:745915 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745721:745919 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745721:745919 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745721:745919 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745721:745919 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745722:745914 [2] NCCL INFO comm 0x9ac4ec0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8f20313144f70574 - Init COMPLETE +ip-26-0-169-207:745720:745913 [0] NCCL INFO comm 0x9315770 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8f20313144f70574 - Init COMPLETE +ip-26-0-169-207:745724:745918 [4] NCCL INFO comm 0xa5b25e0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8f20313144f70574 - Init COMPLETE +ip-26-0-169-207:745726:745920 [6] NCCL INFO comm 0x9c8e160 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8f20313144f70574 - Init COMPLETE +ip-26-0-169-207:745723:745916 [3] NCCL INFO comm 0x9254c00 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8f20313144f70574 - Init COMPLETE +ip-26-0-169-207:745727:745917 [7] NCCL INFO comm 0xa5c85c0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8f20313144f70574 - Init COMPLETE +ip-26-0-169-207:745725:745915 [5] NCCL INFO comm 0x869bf70 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8f20313144f70574 - Init COMPLETE +ip-26-0-169-207:745721:745919 [1] NCCL INFO comm 0x89b3ac0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8f20313144f70574 - Init COMPLETE +ip-26-0-169-207:745720:745946 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:745723:745950 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:745722:745947 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:745721:745948 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:745725:745952 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:745727:745951 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:745724:745949 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:745726:745953 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:745721:745948 [1] NCCL INFO comm 0x89c8050 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf360fb6f47916e05 - Init START +ip-26-0-169-207:745727:745951 [7] NCCL INFO comm 0xa5dcc40 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf360fb6f47916e05 - Init START +ip-26-0-169-207:745720:745946 [0] NCCL INFO comm 0x9329d00 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf360fb6f47916e05 - Init START +ip-26-0-169-207:745724:745949 [4] NCCL INFO comm 0xa5c6d40 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf360fb6f47916e05 - Init START +ip-26-0-169-207:745722:745947 [2] NCCL INFO comm 0x9ad9620 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf360fb6f47916e05 - Init START +ip-26-0-169-207:745726:745953 [6] NCCL INFO comm 0x9ca28c0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf360fb6f47916e05 - Init START +ip-26-0-169-207:745723:745950 [3] NCCL INFO comm 0x92691d0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf360fb6f47916e05 - Init START +ip-26-0-169-207:745725:745952 [5] NCCL INFO comm 0x86b04a0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf360fb6f47916e05 - Init START +ip-26-0-169-207:745727:745951 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745726:745953 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745723:745950 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745724:745949 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745722:745947 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745725:745952 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745721:745948 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745720:745946 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:745724:745949 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745724:745949 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:745725:745952 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745725:745952 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:745726:745953 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745726:745953 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:745722:745947 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:745722:745947 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:745723:745950 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:745723:745950 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:745727:745951 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745727:745951 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:745720:745946 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:745721:745948 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:745721:745948 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:745721:745948 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:745721:745948 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745722:745947 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:745722:745947 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745724:745949 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:745724:745949 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745723:745950 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:745723:745950 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745726:745953 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745726:745953 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:745727:745951 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:745720:745946 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:745720:745946 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745727:745951 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745725:745952 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:745725:745952 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Connected all rings +ip-26-0-169-207:745725:745952 [5] NCCL INFO Connected all rings +ip-26-0-169-207:745721:745948 [1] NCCL INFO Connected all rings +ip-26-0-169-207:745723:745950 [3] NCCL INFO Connected all rings +ip-26-0-169-207:745722:745947 [2] NCCL INFO Connected all rings +ip-26-0-169-207:745720:745946 [0] NCCL INFO Connected all rings +ip-26-0-169-207:745726:745953 [6] NCCL INFO Connected all rings +ip-26-0-169-207:745727:745951 [7] NCCL INFO Connected all rings +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:745951 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745725:745952 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745724:745949 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745721:745948 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745722:745947 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745726:745953 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745723:745950 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:745720:745946 [0] NCCL INFO Connected all trees +ip-26-0-169-207:745720:745946 [0] NCCL INFO NVLS comm 0x9329d00 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745727:745951 [7] NCCL INFO Connected all trees +ip-26-0-169-207:745727:745951 [7] NCCL INFO NVLS comm 0xa5dcc40 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745725:745952 [5] NCCL INFO Connected all trees +ip-26-0-169-207:745725:745952 [5] NCCL INFO NVLS comm 0x86b04a0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745726:745953 [6] NCCL INFO Connected all trees +ip-26-0-169-207:745726:745953 [6] NCCL INFO NVLS comm 0x9ca28c0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745723:745950 [3] NCCL INFO Connected all trees +ip-26-0-169-207:745723:745950 [3] NCCL INFO NVLS comm 0x92691d0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745721:745948 [1] NCCL INFO Connected all trees +ip-26-0-169-207:745721:745948 [1] NCCL INFO NVLS comm 0x89c8050 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745722:745947 [2] NCCL INFO Connected all trees +ip-26-0-169-207:745722:745947 [2] NCCL INFO NVLS comm 0x9ad9620 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745724:745949 [4] NCCL INFO Connected all trees +ip-26-0-169-207:745724:745949 [4] NCCL INFO NVLS comm 0xa5c6d40 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:745725:745952 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745725:745952 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745725:745952 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745725:745952 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745723:745950 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745723:745950 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745723:745950 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745723:745950 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745724:745949 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745724:745949 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745724:745949 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745722:745947 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745724:745949 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745722:745947 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745722:745947 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745722:745947 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745726:745953 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745726:745953 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745726:745953 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745726:745953 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745720:745946 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745720:745946 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745720:745946 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745720:745946 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745721:745948 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745721:745948 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745721:745948 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745721:745948 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745727:745951 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745727:745951 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745727:745951 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:745727:745951 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745724:745949 [4] NCCL INFO comm 0xa5c6d40 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf360fb6f47916e05 - Init COMPLETE +ip-26-0-169-207:745722:745947 [2] NCCL INFO comm 0x9ad9620 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf360fb6f47916e05 - Init COMPLETE +ip-26-0-169-207:745720:745946 [0] NCCL INFO comm 0x9329d00 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf360fb6f47916e05 - Init COMPLETE +ip-26-0-169-207:745726:745953 [6] NCCL INFO comm 0x9ca28c0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf360fb6f47916e05 - Init COMPLETE +ip-26-0-169-207:745723:745950 [3] NCCL INFO comm 0x92691d0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf360fb6f47916e05 - Init COMPLETE +ip-26-0-169-207:745721:745948 [1] NCCL INFO comm 0x89c8050 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf360fb6f47916e05 - Init COMPLETE +ip-26-0-169-207:745727:745951 [7] NCCL INFO comm 0xa5dcc40 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf360fb6f47916e05 - Init COMPLETE +ip-26-0-169-207:745725:745952 [5] NCCL INFO comm 0x86b04a0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf360fb6f47916e05 - Init COMPLETE +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config: +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config(general=GeneralArgs(project='debug', +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: run='1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k', +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: step=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: consumed_train_samples=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ignore_sanity_checks=True), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: parallelism=ParallelismArgs(dp=4, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp=1, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp=2, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp_engine=, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_mode=, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_linear_async_communication=True, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: recompute_layer=False, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_recompute_allgather=True, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: expert_parallel_size=1), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: init_method=RandomInit(std=0.02), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: dtype=torch.bfloat16, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: make_vocab_size_divisible_by=1, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ddp_bucket_cap_mb=25), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_revision=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_max_length=None), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoint_interval=10000, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_initial_state=False, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_final_state=False, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: resume_checkpoint_path=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints_path_is_shared_file_system=False), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: logging=LoggingArgs(log_level='info', +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: log_level_replica='info', +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration_step_info_interval=1), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokens=TokensArgs(sequence_length=2048, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: train_steps=100, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: micro_batch_size=2, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: batch_accumulation_per_replica=256, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: val_check_interval=100, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_val_batches=0, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_test_batches=0), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta1=0.9, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta2=0.95, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: torch_adam_is_fused=True, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: name='adamW'), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: zero_stage=1, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: weight_decay=0.01, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: clip_grad=1.0, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: accumulate_grad_in_fp32=True, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_steps=2, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_style='linear', +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_style='cosine', +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_steps=13, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_starting_step=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: min_decay_lr=1e-05)), +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: start_training_step=1, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data=DataArgs(dataset=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_loading_workers=1))], +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: profiler=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lighteval=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: s3_upload=None) +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Model Config: +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: LlamaConfig(bos_token_id=0, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768) +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Building model.. +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Initialize RoPE Theta = 10000.0 +12/13/2024 15:10:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:745724:746006 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:745723:746007 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:745721:746008 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:745727:746010 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:745726:746005 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:745720:746003 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:745722:746004 [2] NCCL INFO comm 0xb8e9db0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf0914f1e4b189c16 - Init START +ip-26-0-169-207:745723:746007 [3] NCCL INFO comm 0xb076170 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf0914f1e4b189c16 - Init START +ip-26-0-169-207:745722:746004 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745727:746010 [7] NCCL INFO comm 0xc3ea8b0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xee1ebfbcfd117ffb - Init START +ip-26-0-169-207:745726:746005 [6] NCCL INFO comm 0xbaad2e0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xee1ebfbcfd117ffb - Init START +ip-26-0-169-207:745725:746009 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:745726:746005 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745720:746003 [0] NCCL INFO comm 0xb137720 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb516ac0edbf3df80 - Init START +ip-26-0-169-207:745721:746008 [1] NCCL INFO comm 0xa7d3eb0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb516ac0edbf3df80 - Init START +ip-26-0-169-207:745720:746003 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:745724:746006 [4] NCCL INFO comm 0xc3d1630 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xde17fe9026a046d6 - Init START +ip-26-0-169-207:745725:746009 [5] NCCL INFO comm 0xa4bcef0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde17fe9026a046d6 - Init START +ip-26-0-169-207:745724:746006 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745723:746007 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745721:746008 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745727:746010 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745725:746009 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745721:746008 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:745725:746009 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:745721:746008 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:745721:746008 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:746003 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:745720:746003 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745723:746007 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:745724:746006 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:745725:746009 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:745725:746009 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745724:746006 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:745724:746006 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745727:746010 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:745723:746007 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:745723:746007 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745722:746004 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:745722:746004 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:745727:746010 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:745727:746010 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:745726:746005 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746009 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745721:746008 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745722:746004 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745726:746005 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745723:746007 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:745720:746003 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745727:746010 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:745724:746006 [4] NCCL INFO Connected all rings +ip-26-0-169-207:745724:746006 [4] NCCL INFO Connected all trees +ip-26-0-169-207:745724:746006 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745724:746006 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:745724:746006 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745725:746009 [5] NCCL INFO Connected all rings +ip-26-0-169-207:745725:746009 [5] NCCL INFO Connected all trees +ip-26-0-169-207:745725:746009 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745725:746009 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:745725:746009 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745725:746009 [5] NCCL INFO comm 0xa4bcef0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde17fe9026a046d6 - Init COMPLETE +ip-26-0-169-207:745724:746006 [4] NCCL INFO comm 0xc3d1630 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xde17fe9026a046d6 - Init COMPLETE +ip-26-0-169-207:745724:746023 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:745724:746023 [4] NCCL INFO comm 0xc3e4610 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x60b263086764f470 - Init START +ip-26-0-169-207:745724:746023 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:745725:746025 [5] NCCL INFO comm 0xa4cfbb0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5a41379fce9ca813 - Init START +ip-26-0-169-207:745725:746025 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745723:746007 [3] NCCL INFO Connected all rings +ip-26-0-169-207:745723:746007 [3] NCCL INFO Connected all trees +ip-26-0-169-207:745723:746007 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745723:746007 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:745723:746007 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745720:746003 [0] NCCL INFO Connected all rings +ip-26-0-169-207:745720:746003 [0] NCCL INFO Connected all trees +ip-26-0-169-207:745720:746003 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745720:746003 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:745720:746003 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745721:746008 [1] NCCL INFO Connected all rings +ip-26-0-169-207:745721:746008 [1] NCCL INFO Connected all trees +ip-26-0-169-207:745721:746008 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745721:746008 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:745721:746008 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745722:746004 [2] NCCL INFO Connected all rings +ip-26-0-169-207:745722:746004 [2] NCCL INFO Connected all trees +ip-26-0-169-207:745722:746004 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745722:746004 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:745722:746004 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745721:746008 [1] NCCL INFO comm 0xa7d3eb0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb516ac0edbf3df80 - Init COMPLETE +ip-26-0-169-207:745723:746007 [3] NCCL INFO comm 0xb076170 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf0914f1e4b189c16 - Init COMPLETE +ip-26-0-169-207:745722:746004 [2] NCCL INFO comm 0xb8e9db0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf0914f1e4b189c16 - Init COMPLETE +ip-26-0-169-207:745720:746003 [0] NCCL INFO comm 0xb137720 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb516ac0edbf3df80 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:745721:746032 [1] NCCL INFO comm 0xa7e6b70 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6415008a55c98ac9 - Init START +ip-26-0-169-207:745721:746032 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:745720:746034 [0] NCCL INFO comm 0xb14a810 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x798694772e1219f3 - Init START +ip-26-0-169-207:745720:746034 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:745722:746036 [2] NCCL INFO comm 0xb8fcea0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x28d9dd47697927d1 - Init START +ip-26-0-169-207:745722:746036 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:745723:746038 [3] NCCL INFO comm 0xb088e30 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x179795930ca5553a - Init START +ip-26-0-169-207:745723:746038 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745727:746010 [7] NCCL INFO Connected all rings +ip-26-0-169-207:745727:746010 [7] NCCL INFO Connected all trees +ip-26-0-169-207:745727:746010 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745727:746010 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:745727:746010 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745726:746005 [6] NCCL INFO Connected all rings +ip-26-0-169-207:745726:746005 [6] NCCL INFO Connected all trees +ip-26-0-169-207:745726:746005 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745726:746005 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:745726:746005 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745727:746010 [7] NCCL INFO comm 0xc3ea8b0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xee1ebfbcfd117ffb - Init COMPLETE +ip-26-0-169-207:745726:746005 [6] NCCL INFO comm 0xbaad2e0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xee1ebfbcfd117ffb - Init COMPLETE +ip-26-0-169-207:745726:746042 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:745726:746042 [6] NCCL INFO comm 0xbac03d0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x16c650cb56aefbcb - Init START +ip-26-0-169-207:745726:746042 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:745727:746044 [7] NCCL INFO comm 0xc3fd570 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x31621c25200783b8 - Init START +ip-26-0-169-207:745727:746044 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:745725:746025 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:745725:746025 [5] NCCL INFO Connected all rings +ip-26-0-169-207:745725:746025 [5] NCCL INFO Connected all trees +ip-26-0-169-207:745725:746025 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745724:746023 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:745724:746023 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:745724:746023 [4] NCCL INFO Connected all rings +ip-26-0-169-207:745724:746023 [4] NCCL INFO Connected all trees +ip-26-0-169-207:745724:746023 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745725:746025 [5] NCCL INFO comm 0xa4cfbb0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5a41379fce9ca813 - Init COMPLETE +ip-26-0-169-207:745724:746023 [4] NCCL INFO comm 0xc3e4610 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x60b263086764f470 - Init COMPLETE +ip-26-0-169-207:745722:746036 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:745722:746036 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:745722:746036 [2] NCCL INFO Connected all rings +ip-26-0-169-207:745722:746036 [2] NCCL INFO Connected all trees +ip-26-0-169-207:745722:746036 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745722:746036 [2] NCCL INFO comm 0xb8fcea0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x28d9dd47697927d1 - Init COMPLETE +ip-26-0-169-207:745721:746032 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:745721:746032 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:745721:746032 [1] NCCL INFO Connected all rings +ip-26-0-169-207:745721:746032 [1] NCCL INFO Connected all trees +ip-26-0-169-207:745721:746032 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745723:746038 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:745723:746038 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:745723:746038 [3] NCCL INFO Connected all rings +ip-26-0-169-207:745723:746038 [3] NCCL INFO Connected all trees +ip-26-0-169-207:745723:746038 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745720:746034 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:745720:746034 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:745720:746034 [0] NCCL INFO Connected all rings +ip-26-0-169-207:745720:746034 [0] NCCL INFO Connected all trees +ip-26-0-169-207:745720:746034 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745721:746032 [1] NCCL INFO comm 0xa7e6b70 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6415008a55c98ac9 - Init COMPLETE +ip-26-0-169-207:745723:746038 [3] NCCL INFO comm 0xb088e30 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x179795930ca5553a - Init COMPLETE +ip-26-0-169-207:745727:746044 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:745727:746044 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Connected all rings +ip-26-0-169-207:745726:746042 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745727:746044 [7] NCCL INFO Connected all trees +ip-26-0-169-207:745727:746044 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:745726:746042 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:745726:746042 [6] NCCL INFO Connected all rings +ip-26-0-169-207:745726:746042 [6] NCCL INFO Connected all trees +ip-26-0-169-207:745726:746042 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745720:746034 [0] NCCL INFO comm 0xb14a810 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x798694772e1219f3 - Init COMPLETE +12/13/2024 15:10:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Total number of parameters: 1.04G (1984.26MiB) +12/13/2024 15:10:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +12/13/2024 15:10:48 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +12/13/2024 15:10:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/13/2024 15:10:48 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/13/2024 15:10:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: No checkpoint path provided. +12/13/2024 15:10:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Parametrizing model parameters using StandardParametrizator +ip-26-0-169-207:745722:746063 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:745724:746064 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:745720:746062 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:745725:746068 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:745721:746066 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:745723:746067 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:745727:746044 [7] NCCL INFO comm 0xc3fd570 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x31621c25200783b8 - Init COMPLETE +ip-26-0-169-207:745726:746042 [6] NCCL INFO comm 0xbac03d0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x16c650cb56aefbcb - Init COMPLETE +ip-26-0-169-207:745726:746069 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:745727:746070 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:745724:746064 [4] NCCL INFO comm 0xc3e72c0 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf5e3492264db13d1 - Init START +ip-26-0-169-207:745722:746063 [2] NCCL INFO comm 0xb8ffb50 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf5e3492264db13d1 - Init START +ip-26-0-169-207:745726:746069 [6] NCCL INFO comm 0xbac3080 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf5e3492264db13d1 - Init START +ip-26-0-169-207:745720:746062 [0] NCCL INFO comm 0xb14d4c0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf5e3492264db13d1 - Init START +ip-26-0-169-207:745721:746066 [1] NCCL INFO comm 0xa7e9820 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x920588af71489d37 - Init START +ip-26-0-169-207:745727:746070 [7] NCCL INFO comm 0xc400220 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x920588af71489d37 - Init START +ip-26-0-169-207:745723:746067 [3] NCCL INFO comm 0xb08bae0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x920588af71489d37 - Init START +ip-26-0-169-207:745725:746068 [5] NCCL INFO comm 0xa4d2860 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x920588af71489d37 - Init START +ip-26-0-169-207:745724:746064 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745726:746069 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745722:746063 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745720:746062 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:745723:746067 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745725:746068 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745727:746070 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745721:746066 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:745723:746067 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:745723:746067 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:745727:746070 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745727:746070 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:745720:746062 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:745721:746066 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:745722:746063 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:745722:746063 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:745724:746064 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745724:746064 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:745725:746068 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745725:746068 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:745726:746069 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:745726:746069 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:745725:746068 [5] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:745725:746068 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:745727:746070 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:745727:746070 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:745723:746067 [3] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:745723:746067 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:745721:746066 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:745726:746069 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:745724:746064 [4] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:745726:746069 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745722:746063 [2] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:745724:746064 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:745722:746063 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:745720:746062 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 04/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 04/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 05/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 05/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 06/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 06/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 07/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 07/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 08/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 08/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 09/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 09/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 10/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 10/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 11/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 11/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 12/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 12/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 13/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 13/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 14/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 14/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 15/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 15/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 16/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 16/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 17/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 17/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 18/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 18/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 19/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 19/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 20/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 04/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 21/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 05/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 20/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 04/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 22/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 06/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 05/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 21/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 06/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 22/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 07/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 23/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 23/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 08/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 07/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 09/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 08/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 04/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 10/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 09/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 05/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 11/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 10/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 06/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 12/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 11/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 07/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 13/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 12/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 08/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 14/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 13/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 09/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 15/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745721:746066 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 14/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 10/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 16/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 15/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 11/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 17/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 16/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 12/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 18/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 17/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 13/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 19/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 18/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 14/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 20/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 19/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 15/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 21/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 20/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 16/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 22/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 21/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 17/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 23/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 22/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 23/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 18/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 19/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 20/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 21/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 22/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 23/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Connected all rings +ip-26-0-169-207:745725:746068 [5] NCCL INFO Connected all rings +ip-26-0-169-207:745721:746066 [1] NCCL INFO Connected all rings +ip-26-0-169-207:745727:746070 [7] NCCL INFO Connected all rings +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Connected all rings +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Connected all rings +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 02/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 03/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 04/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745720:746062 [0] NCCL INFO Connected all rings +ip-26-0-169-207:745724:746064 [4] NCCL INFO Connected all rings +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 05/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 06/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 07/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 08/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 09/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 10/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 11/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 12/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 13/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 14/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 15/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 16/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 17/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 18/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 04/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 19/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 05/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 20/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 06/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 21/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 07/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 22/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 08/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745726:746069 [6] NCCL INFO Channel 23/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 09/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 10/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 11/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 12/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 13/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 14/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 15/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 16/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 17/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 18/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 04/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 19/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 05/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 20/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 06/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 21/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 07/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 22/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 08/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745725:746068 [5] NCCL INFO Channel 23/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 09/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 10/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 11/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 12/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 13/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745723:746067 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 14/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 15/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 16/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 17/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 18/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 19/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 20/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 21/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745722:746063 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 22/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745724:746064 [4] NCCL INFO Channel 23/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:745727:746070 [7] NCCL INFO Connected all trees +ip-26-0-169-207:745727:746070 [7] NCCL INFO NVLS comm 0xc400220 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Connected all trees +ip-26-0-169-207:745721:746066 [1] NCCL INFO NVLS comm 0xa7e9820 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:745720:746062 [0] NCCL INFO Connected all trees +ip-26-0-169-207:745720:746062 [0] NCCL INFO NVLS comm 0xb14d4c0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:745726:746069 [6] NCCL INFO Connected all trees +ip-26-0-169-207:745726:746069 [6] NCCL INFO NVLS comm 0xbac3080 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:745725:746068 [5] NCCL INFO Connected all trees +ip-26-0-169-207:745725:746068 [5] NCCL INFO NVLS comm 0xa4d2860 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:745723:746067 [3] NCCL INFO Connected all trees +ip-26-0-169-207:745723:746067 [3] NCCL INFO NVLS comm 0xb08bae0 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:745722:746063 [2] NCCL INFO Connected all trees +ip-26-0-169-207:745722:746063 [2] NCCL INFO NVLS comm 0xb8ffb50 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:745724:746064 [4] NCCL INFO Connected all trees +ip-26-0-169-207:745724:746064 [4] NCCL INFO NVLS comm 0xc3e72c0 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:745721:746066 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745721:746066 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745721:746066 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:745721:746066 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745723:746067 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745723:746067 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745723:746067 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:745723:746067 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745727:746070 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745727:746070 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745727:746070 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:745727:746070 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745725:746068 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745725:746068 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745725:746068 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:745725:746068 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745727:746070 [7] NCCL INFO comm 0xc400220 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x920588af71489d37 - Init COMPLETE +ip-26-0-169-207:745723:746067 [3] NCCL INFO comm 0xb08bae0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x920588af71489d37 - Init COMPLETE +ip-26-0-169-207:745721:746066 [1] NCCL INFO comm 0xa7e9820 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x920588af71489d37 - Init COMPLETE +ip-26-0-169-207:745725:746068 [5] NCCL INFO comm 0xa4d2860 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x920588af71489d37 - Init COMPLETE +ip-26-0-169-207:745726:746069 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745726:746069 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745726:746069 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:745726:746069 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745722:746063 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745722:746063 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745722:746063 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:745722:746063 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745720:746062 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745720:746062 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745724:746064 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:745720:746062 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:745720:746062 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745724:746064 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:745724:746064 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:745724:746064 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:745722:746063 [2] NCCL INFO comm 0xb8ffb50 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf5e3492264db13d1 - Init COMPLETE +ip-26-0-169-207:745720:746062 [0] NCCL INFO comm 0xb14d4c0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf5e3492264db13d1 - Init COMPLETE +ip-26-0-169-207:745724:746064 [4] NCCL INFO comm 0xc3e72c0 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf5e3492264db13d1 - Init COMPLETE +ip-26-0-169-207:745726:746069 [6] NCCL INFO comm 0xbac3080 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf5e3492264db13d1 - Init COMPLETE +12/13/2024 15:10:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Optimizer Building] Using LearningRateForSP as learning rate +12/13/2024 15:10:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] Size of optimizer params per rank: +12/13/2024 15:10:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 0 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 15:10:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 1 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 15:10:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 2 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 15:10:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 3 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 15:10:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/13/2024 15:10:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Using dummy data generator +12/13/2024 15:10:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] There are 1 training stages +12/13/2024 15:10:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Stage Stable Training Stage] start from step 1 +12/13/2024 15:10:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: +12/13/2024 15:10:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Start training] datetime: 2024-12-13 15:10:53.612166 | mbs: 2 | grad_accum: 256 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/13/2024 15:11:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Waiting for 17 seconds +12/13/2024 15:11:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/13/2024 15:11:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3472.48MiB. Peak allocated 5632.00MiB. Peak reserved: 18884.00MiB +12/13/2024 15:11:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3572.67MiB. Peak allocated 6985.16MiB. Peak reserved: 7454.00MiB +12/13/2024 15:11:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 20.1K | tokens_per_sec: 209K | tokens_per_sec_per_gpu: 26.1K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 184 | hardware_tflops_per_gpu: 184 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.64G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 15:11:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4588.85MiB. Peak allocated 6821.14MiB. Peak reserved: 9198.00MiB +12/13/2024 15:11:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4588.97MiB. Peak allocated 8001.47MiB. Peak reserved: 9198.00MiB +12/13/2024 15:11:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 18.7K | tokens_per_sec: 224K | tokens_per_sec_per_gpu: 28.1K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.0003 | model_tflops_per_gpu: 198 | hardware_tflops_per_gpu: 198 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.64G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 15:11:57 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4588.85MiB. Peak allocated 6821.26MiB. Peak reserved: 9198.00MiB +12/13/2024 15:12:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4588.97MiB. Peak allocated 8001.47MiB. Peak reserved: 9198.00MiB +12/13/2024 15:12:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 3 / 100 | consumed_tokens: 12.6M | elapsed_time_per_iteration_ms: 18.7K | tokens_per_sec: 224K | tokens_per_sec_per_gpu: 28K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.000296 | model_tflops_per_gpu: 198 | hardware_tflops_per_gpu: 198 | grad_norm: 0.424 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.64G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 15:12:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/13/2024 15:12:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: | -------- | -------------------------------------------------------------- | ----- | ------- | --- | ----------- | ---- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | --------------------------------------------------------------------------------------------------------- | --------------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/13/2024 15:12:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: | 13417921 | 1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k | 1 | 2048 | 2 | 256 | 2048 | 197.53 | 197.53 | 28032.45 | 460.58 | 263.72 | 262.75 | 459.05 | 265.17 | 263.89 | 6.66 | 8.98 | 4 | 1 | 2 | | TensorParallelLinearMode.REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 8 | 2048 | 32768 | True | torch.bfloat16 | 1 | 25 | True | 1.04G | 520M | +12/13/2024 15:12:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final.csv +12/13/2024 15:12:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Acquired lock for CSV file: benchmark/results/bench_final.csv +slurmstepd: error: *** STEP 13417921.0 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T15:12:32 DUE TO TIME LIMIT *** +slurmstepd: error: *** JOB 13417921 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T15:12:32 DUE TO TIME LIMIT *** +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-13 15:12:32,299] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-13 15:12:32,299] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745720 closing signal SIGTERM +[2024-12-13 15:12:32,299] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745721 closing signal SIGTERM +[2024-12-13 15:12:32,299] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745722 closing signal SIGTERM +[2024-12-13 15:12:32,299] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745723 closing signal SIGTERM +[2024-12-13 15:12:32,302] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745724 closing signal SIGTERM +[2024-12-13 15:12:32,303] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745725 closing signal SIGTERM +[2024-12-13 15:12:32,303] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745726 closing signal SIGTERM +[2024-12-13 15:12:32,303] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745727 closing signal SIGTERM +[2024-12-13 15:12:32,304] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745720 closing signal SIGTERM +[2024-12-13 15:12:32,304] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745721 closing signal SIGTERM +[2024-12-13 15:12:32,304] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745722 closing signal SIGTERM +[2024-12-13 15:12:32,304] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745723 closing signal SIGTERM +[2024-12-13 15:12:32,304] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745724 closing signal SIGTERM +[2024-12-13 15:12:32,304] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745725 closing signal SIGTERM +[2024-12-13 15:12:32,304] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745726 closing signal SIGTERM +[2024-12-13 15:12:32,304] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 745727 closing signal SIGTERM diff --git a/logs/13417991-bench_stress_test.out b/logs/13417991-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..e612b9ec31a620f417dce2d6e458aa8fa5561500 --- /dev/null +++ b/logs/13417991-bench_stress_test.out @@ -0,0 +1,3768 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13417991 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 16:39:03,777] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 16:39:03,777] torch.distributed.run: [WARNING] +[2024-12-13 16:39:03,777] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 16:39:03,777] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 16:39:03,777] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-169-207:782346:782346 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:782346:782346 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:782346:782346 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:782346:782346 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:782346:782346 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:782351:782351 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:782347:782347 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:782351:782351 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:782347:782347 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:782351:782351 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:782347:782347 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:782352:782352 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:782349:782349 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:782347:782347 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:782347:782347 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:782352:782352 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:782349:782349 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:782351:782351 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:782351:782351 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:782350:782350 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:782350:782350 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:782353:782353 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:782348:782348 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:782352:782352 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:782353:782353 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:782348:782348 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:782349:782349 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:782352:782352 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:782352:782352 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:782350:782350 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:782349:782349 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:782349:782349 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:782353:782353 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:782348:782348 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:782350:782350 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:782350:782350 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:782353:782353 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:782348:782348 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:782353:782353 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:782348:782348 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/162 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:782353:782473 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:782353:782473 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:782350:782471 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:782350:782471 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:782346:782467 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:782346:782467 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:782352:782472 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:782352:782472 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:782351:782468 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:782351:782468 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:782347:782469 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:782347:782469 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:782348:782474 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:782348:782474 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:782349:782470 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:782349:782470 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:782350:782471 [4] NCCL INFO comm 0x945cfc0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb6c60c6633ea5ef1 - Init START +ip-26-0-169-207:782348:782474 [2] NCCL INFO comm 0x95a8920 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb6c60c6633ea5ef1 - Init START +ip-26-0-169-207:782347:782469 [1] NCCL INFO comm 0x910f8a0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb6c60c6633ea5ef1 - Init START +ip-26-0-169-207:782351:782468 [5] NCCL INFO comm 0x91f9f30 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb6c60c6633ea5ef1 - Init START +ip-26-0-169-207:782349:782470 [3] NCCL INFO comm 0x84462b0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb6c60c6633ea5ef1 - Init START +ip-26-0-169-207:782352:782472 [6] NCCL INFO comm 0x8752fe0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb6c60c6633ea5ef1 - Init START +ip-26-0-169-207:782346:782467 [0] NCCL INFO comm 0x9736ad0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb6c60c6633ea5ef1 - Init START +ip-26-0-169-207:782353:782473 [7] NCCL INFO comm 0x95767b0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb6c60c6633ea5ef1 - Init START +ip-26-0-169-207:782352:782472 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782348:782474 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782347:782469 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782351:782468 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:782349:782470 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782350:782471 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782350:782471 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:782352:782472 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782352:782472 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:782347:782469 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:782347:782469 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:782353:782473 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782353:782473 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:782348:782474 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:782348:782474 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:782349:782470 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:782349:782470 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:782351:782468 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:782351:782468 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:782346:782467 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782467 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:782346:782467 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782353:782473 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:782352:782472 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:782353:782473 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782351:782468 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:782352:782472 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782351:782468 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782349:782470 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:782349:782470 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782350:782471 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:782350:782471 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782347:782469 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:782347:782469 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782348:782474 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:782348:782474 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Connected all rings +ip-26-0-169-207:782350:782471 [4] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782473 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782469 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782352:782472 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782471 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782474 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782470 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782351:782468 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782346:782467 [0] NCCL INFO Connected all trees +ip-26-0-169-207:782346:782467 [0] NCCL INFO NVLS comm 0x9736ad0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782353:782473 [7] NCCL INFO Connected all trees +ip-26-0-169-207:782353:782473 [7] NCCL INFO NVLS comm 0x95767b0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782349:782470 [3] NCCL INFO Connected all trees +ip-26-0-169-207:782349:782470 [3] NCCL INFO NVLS comm 0x84462b0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782347:782469 [1] NCCL INFO Connected all trees +ip-26-0-169-207:782347:782469 [1] NCCL INFO NVLS comm 0x910f8a0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782350:782471 [4] NCCL INFO Connected all trees +ip-26-0-169-207:782350:782471 [4] NCCL INFO NVLS comm 0x945cfc0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782351:782468 [5] NCCL INFO Connected all trees +ip-26-0-169-207:782351:782468 [5] NCCL INFO NVLS comm 0x91f9f30 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782348:782474 [2] NCCL INFO Connected all trees +ip-26-0-169-207:782348:782474 [2] NCCL INFO NVLS comm 0x95a8920 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782352:782472 [6] NCCL INFO Connected all trees +ip-26-0-169-207:782352:782472 [6] NCCL INFO NVLS comm 0x8752fe0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782349:782470 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782349:782470 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782349:782470 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782349:782470 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782348:782474 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782348:782474 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782348:782474 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782348:782474 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782347:782469 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782347:782469 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782347:782469 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782347:782469 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782353:782473 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782353:782473 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782353:782473 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782353:782473 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782351:782468 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782351:782468 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782351:782468 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782351:782468 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782352:782472 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782352:782472 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782352:782472 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782352:782472 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782346:782467 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782346:782467 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782350:782471 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782346:782467 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782346:782467 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782350:782471 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782350:782471 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782350:782471 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782350:782471 [4] NCCL INFO comm 0x945cfc0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb6c60c6633ea5ef1 - Init COMPLETE +ip-26-0-169-207:782352:782472 [6] NCCL INFO comm 0x8752fe0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb6c60c6633ea5ef1 - Init COMPLETE +ip-26-0-169-207:782348:782474 [2] NCCL INFO comm 0x95a8920 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb6c60c6633ea5ef1 - Init COMPLETE +ip-26-0-169-207:782346:782467 [0] NCCL INFO comm 0x9736ad0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb6c60c6633ea5ef1 - Init COMPLETE +ip-26-0-169-207:782353:782473 [7] NCCL INFO comm 0x95767b0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb6c60c6633ea5ef1 - Init COMPLETE +ip-26-0-169-207:782349:782470 [3] NCCL INFO comm 0x84462b0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb6c60c6633ea5ef1 - Init COMPLETE +ip-26-0-169-207:782347:782469 [1] NCCL INFO comm 0x910f8a0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb6c60c6633ea5ef1 - Init COMPLETE +ip-26-0-169-207:782351:782468 [5] NCCL INFO comm 0x91f9f30 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb6c60c6633ea5ef1 - Init COMPLETE +ip-26-0-169-207:782352:782540 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:782353:782542 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:782349:782545 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:782350:782544 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:782348:782543 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:782346:782539 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:782351:782541 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:782347:782546 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:782350:782544 [4] NCCL INFO comm 0x9686b20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5fa179aacd95e2c - Init START +ip-26-0-169-207:782348:782543 [2] NCCL INFO comm 0x97d2ac0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5fa179aacd95e2c - Init START +ip-26-0-169-207:782349:782545 [3] NCCL INFO comm 0x8670130 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5fa179aacd95e2c - Init START +ip-26-0-169-207:782347:782546 [1] NCCL INFO comm 0x9339df0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5fa179aacd95e2c - Init START +ip-26-0-169-207:782346:782539 [0] NCCL INFO comm 0x99605f0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5fa179aacd95e2c - Init START +ip-26-0-169-207:782352:782540 [6] NCCL INFO comm 0x897c9c0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5fa179aacd95e2c - Init START +ip-26-0-169-207:782351:782541 [5] NCCL INFO comm 0x94241c0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5fa179aacd95e2c - Init START +ip-26-0-169-207:782353:782542 [7] NCCL INFO comm 0x97a0330 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5fa179aacd95e2c - Init START +ip-26-0-169-207:782346:782539 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:782347:782546 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782353:782542 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782351:782541 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782350:782544 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782349:782545 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782352:782540 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782348:782543 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782350:782544 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782350:782544 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:782352:782540 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782352:782540 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:782346:782539 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:782353:782542 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782353:782542 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:782349:782545 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:782349:782545 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:782348:782543 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:782348:782543 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:782351:782541 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782351:782541 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:782347:782546 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:782347:782546 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:782347:782546 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:782348:782543 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:782347:782546 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782348:782543 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782349:782545 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:782349:782545 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782351:782541 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:782352:782540 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:782351:782541 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782352:782540 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782353:782542 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:782353:782542 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782350:782544 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782350:782544 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782539 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:782346:782539 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Connected all rings +ip-26-0-169-207:782346:782539 [0] NCCL INFO Connected all rings +ip-26-0-169-207:782347:782546 [1] NCCL INFO Connected all rings +ip-26-0-169-207:782348:782543 [2] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782542 [7] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Connected all rings +ip-26-0-169-207:782352:782540 [6] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782353:782542 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782545 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782540 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782546 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782544 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782543 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782351:782541 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782346:782539 [0] NCCL INFO Connected all trees +ip-26-0-169-207:782346:782539 [0] NCCL INFO NVLS comm 0x99605f0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782353:782542 [7] NCCL INFO Connected all trees +ip-26-0-169-207:782353:782542 [7] NCCL INFO NVLS comm 0x97a0330 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782351:782541 [5] NCCL INFO Connected all trees +ip-26-0-169-207:782351:782541 [5] NCCL INFO NVLS comm 0x94241c0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782347:782546 [1] NCCL INFO Connected all trees +ip-26-0-169-207:782347:782546 [1] NCCL INFO NVLS comm 0x9339df0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782349:782545 [3] NCCL INFO Connected all trees +ip-26-0-169-207:782349:782545 [3] NCCL INFO NVLS comm 0x8670130 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782352:782540 [6] NCCL INFO Connected all trees +ip-26-0-169-207:782352:782540 [6] NCCL INFO NVLS comm 0x897c9c0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782350:782544 [4] NCCL INFO Connected all trees +ip-26-0-169-207:782350:782544 [4] NCCL INFO NVLS comm 0x9686b20 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782348:782543 [2] NCCL INFO Connected all trees +ip-26-0-169-207:782348:782543 [2] NCCL INFO NVLS comm 0x97d2ac0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782349:782545 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782349:782545 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782349:782545 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782349:782545 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782348:782543 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782352:782540 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782348:782543 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782352:782540 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782348:782543 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782348:782543 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782352:782540 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782352:782540 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782353:782542 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782353:782542 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782353:782542 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782353:782542 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782350:782544 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782351:782541 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782351:782541 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782350:782544 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782351:782541 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782351:782541 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782350:782544 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782350:782544 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782347:782546 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782347:782546 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782347:782546 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782347:782546 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782346:782539 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782346:782539 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782346:782539 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782346:782539 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782353:782542 [7] NCCL INFO comm 0x97a0330 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5fa179aacd95e2c - Init COMPLETE +ip-26-0-169-207:782347:782546 [1] NCCL INFO comm 0x9339df0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5fa179aacd95e2c - Init COMPLETE +ip-26-0-169-207:782349:782545 [3] NCCL INFO comm 0x8670130 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5fa179aacd95e2c - Init COMPLETE +ip-26-0-169-207:782351:782541 [5] NCCL INFO comm 0x94241c0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5fa179aacd95e2c - Init COMPLETE +ip-26-0-169-207:782348:782543 [2] NCCL INFO comm 0x97d2ac0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5fa179aacd95e2c - Init COMPLETE +ip-26-0-169-207:782352:782540 [6] NCCL INFO comm 0x897c9c0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5fa179aacd95e2c - Init COMPLETE +ip-26-0-169-207:782346:782539 [0] NCCL INFO comm 0x99605f0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5fa179aacd95e2c - Init COMPLETE +ip-26-0-169-207:782350:782544 [4] NCCL INFO comm 0x9686b20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5fa179aacd95e2c - Init COMPLETE +ip-26-0-169-207:782346:782572 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:782351:782576 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:782348:782573 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:782349:782575 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:782347:782574 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:782352:782578 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:782353:782577 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:782350:782579 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:782349:782575 [3] NCCL INFO comm 0x86846c0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7e5ba44ce088b27f - Init START +ip-26-0-169-207:782348:782573 [2] NCCL INFO comm 0x97e7210 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7e5ba44ce088b27f - Init START +ip-26-0-169-207:782353:782577 [7] NCCL INFO comm 0x97b4810 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7e5ba44ce088b27f - Init START +ip-26-0-169-207:782352:782578 [6] NCCL INFO comm 0x89910e0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7e5ba44ce088b27f - Init START +ip-26-0-169-207:782351:782576 [5] NCCL INFO comm 0x9438710 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7e5ba44ce088b27f - Init START +ip-26-0-169-207:782346:782572 [0] NCCL INFO comm 0x9974c60 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7e5ba44ce088b27f - Init START +ip-26-0-169-207:782347:782574 [1] NCCL INFO comm 0x934e380 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7e5ba44ce088b27f - Init START +ip-26-0-169-207:782350:782579 [4] NCCL INFO comm 0x969b1f0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7e5ba44ce088b27f - Init START +ip-26-0-169-207:782348:782573 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782349:782575 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782347:782574 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782353:782577 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782350:782579 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782351:782576 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782352:782578 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782346:782572 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:782350:782579 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782350:782579 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:782347:782574 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:782347:782574 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:782351:782576 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782351:782576 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:782346:782572 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:782352:782578 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782352:782578 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:782349:782575 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:782349:782575 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:782353:782577 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782353:782577 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:782348:782573 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:782348:782573 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:782348:782573 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:782348:782573 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782349:782575 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:782349:782575 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782352:782578 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:782350:782579 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:782351:782576 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:782353:782577 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:782352:782578 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782351:782576 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782353:782577 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782350:782579 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:782346:782572 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782347:782574 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:782347:782574 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782346:782572 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Connected all rings +ip-26-0-169-207:782351:782576 [5] NCCL INFO Connected all rings +ip-26-0-169-207:782350:782579 [4] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782577 [7] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Connected all rings +ip-26-0-169-207:782346:782572 [0] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782352:782578 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782350:782579 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782349:782575 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782576 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782574 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782348:782573 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:782353:782577 [7] NCCL INFO Connected all trees +ip-26-0-169-207:782353:782577 [7] NCCL INFO NVLS comm 0x97b4810 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782346:782572 [0] NCCL INFO Connected all trees +ip-26-0-169-207:782346:782572 [0] NCCL INFO NVLS comm 0x9974c60 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782352:782578 [6] NCCL INFO Connected all trees +ip-26-0-169-207:782352:782578 [6] NCCL INFO NVLS comm 0x89910e0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782351:782576 [5] NCCL INFO Connected all trees +ip-26-0-169-207:782351:782576 [5] NCCL INFO NVLS comm 0x9438710 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782350:782579 [4] NCCL INFO Connected all trees +ip-26-0-169-207:782350:782579 [4] NCCL INFO NVLS comm 0x969b1f0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782347:782574 [1] NCCL INFO Connected all trees +ip-26-0-169-207:782347:782574 [1] NCCL INFO NVLS comm 0x934e380 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782349:782575 [3] NCCL INFO Connected all trees +ip-26-0-169-207:782349:782575 [3] NCCL INFO NVLS comm 0x86846c0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782348:782573 [2] NCCL INFO Connected all trees +ip-26-0-169-207:782348:782573 [2] NCCL INFO NVLS comm 0x97e7210 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:782349:782575 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782349:782575 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782349:782575 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782349:782575 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782351:782576 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782352:782578 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782351:782576 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782352:782578 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782352:782578 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782352:782578 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782351:782576 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782351:782576 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782348:782573 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782348:782573 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782350:782579 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782348:782573 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782348:782573 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782347:782574 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782350:782579 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782347:782574 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782350:782579 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782347:782574 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782350:782579 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782347:782574 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782346:782572 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782346:782572 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782346:782572 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782346:782572 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782353:782577 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782353:782577 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782353:782577 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:782353:782577 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782352:782578 [6] NCCL INFO comm 0x89910e0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7e5ba44ce088b27f - Init COMPLETE +ip-26-0-169-207:782348:782573 [2] NCCL INFO comm 0x97e7210 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7e5ba44ce088b27f - Init COMPLETE +ip-26-0-169-207:782350:782579 [4] NCCL INFO comm 0x969b1f0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7e5ba44ce088b27f - Init COMPLETE +ip-26-0-169-207:782346:782572 [0] NCCL INFO comm 0x9974c60 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7e5ba44ce088b27f - Init COMPLETE +ip-26-0-169-207:782347:782574 [1] NCCL INFO comm 0x934e380 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7e5ba44ce088b27f - Init COMPLETE +ip-26-0-169-207:782349:782575 [3] NCCL INFO comm 0x86846c0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7e5ba44ce088b27f - Init COMPLETE +ip-26-0-169-207:782351:782576 [5] NCCL INFO comm 0x9438710 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7e5ba44ce088b27f - Init COMPLETE +ip-26-0-169-207:782353:782577 [7] NCCL INFO comm 0x97b4810 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7e5ba44ce088b27f - Init COMPLETE +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config: +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config(general=GeneralArgs(project='debug', +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: run='1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k', +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: step=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: consumed_train_samples=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ignore_sanity_checks=True), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: parallelism=ParallelismArgs(dp=4, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp=1, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp=2, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp_engine=, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_mode=, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_linear_async_communication=True, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: recompute_layer=False, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_recompute_allgather=True, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: expert_parallel_size=1), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: init_method=RandomInit(std=0.02), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: dtype=torch.bfloat16, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: make_vocab_size_divisible_by=1, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ddp_bucket_cap_mb=25), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_revision=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_max_length=None), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoint_interval=10000, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_initial_state=False, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_final_state=False, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: resume_checkpoint_path=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints_path_is_shared_file_system=False), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: logging=LoggingArgs(log_level='info', +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: log_level_replica='info', +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration_step_info_interval=1), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokens=TokensArgs(sequence_length=2048, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: train_steps=100, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: micro_batch_size=2, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: batch_accumulation_per_replica=256, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: val_check_interval=100, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_val_batches=0, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_test_batches=0), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta1=0.9, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta2=0.95, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: torch_adam_is_fused=True, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: name='adamW'), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: zero_stage=1, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: weight_decay=0.01, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: clip_grad=1.0, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: accumulate_grad_in_fp32=True, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_steps=2, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_style='linear', +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_style='cosine', +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_steps=13, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_starting_step=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: min_decay_lr=1e-05)), +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: start_training_step=1, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data=DataArgs(dataset=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_loading_workers=1))], +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: profiler=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lighteval=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: s3_upload=None) +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Model Config: +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: LlamaConfig(bos_token_id=0, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768) +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Building model.. +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Initialize RoPE Theta = 10000.0 +12/13/2024 16:39:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:782350:782628 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:782347:782630 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:782352:782627 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:782348:782629 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:782349:782633 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:782353:782631 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:782351:782632 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:782346:782626 [0] NCCL INFO comm 0xb779ff0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9d5614c730dd244b - Init START +ip-26-0-169-207:782347:782630 [1] NCCL INFO comm 0xb14edb0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9d5614c730dd244b - Init START +ip-26-0-169-207:782353:782631 [7] NCCL INFO comm 0xb5b5520 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfe3ee127b88530f8 - Init START +ip-26-0-169-207:782352:782627 [6] NCCL INFO comm 0xa7934f0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfe3ee127b88530f8 - Init START +ip-26-0-169-207:782350:782628 [4] NCCL INFO comm 0xb49bbb0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x296dfc18a43477d9 - Init START +ip-26-0-169-207:782351:782632 [5] NCCL INFO comm 0xb23dff0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x296dfc18a43477d9 - Init START +ip-26-0-169-207:782346:782626 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:782348:782629 [2] NCCL INFO comm 0xb5e9530 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x97a07c9a493552cd - Init START +ip-26-0-169-207:782349:782633 [3] NCCL INFO comm 0xa4858b0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x97a07c9a493552cd - Init START +ip-26-0-169-207:782352:782627 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782350:782628 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782348:782629 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782349:782633 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782347:782630 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782353:782631 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782351:782632 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782347:782630 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:782349:782633 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:782353:782631 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:782353:782631 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:782353:782631 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:782352:782627 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:782347:782630 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:782346:782626 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782347:782630 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782351:782632 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782351:782632 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:782351:782632 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:782350:782628 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:782349:782633 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:782349:782633 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:782348:782629 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782352:782627 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782632 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782353:782631 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782633 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782346:782626 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782348:782629 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782630 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782628 [4] NCCL INFO Connected all rings +ip-26-0-169-207:782350:782628 [4] NCCL INFO Connected all trees +ip-26-0-169-207:782350:782628 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782350:782628 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:782350:782628 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782351:782632 [5] NCCL INFO Connected all rings +ip-26-0-169-207:782351:782632 [5] NCCL INFO Connected all trees +ip-26-0-169-207:782351:782632 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782351:782632 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:782351:782632 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782352:782627 [6] NCCL INFO Connected all rings +ip-26-0-169-207:782352:782627 [6] NCCL INFO Connected all trees +ip-26-0-169-207:782352:782627 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782352:782627 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:782352:782627 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782350:782628 [4] NCCL INFO comm 0xb49bbb0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x296dfc18a43477d9 - Init COMPLETE +ip-26-0-169-207:782351:782632 [5] NCCL INFO comm 0xb23dff0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x296dfc18a43477d9 - Init COMPLETE +ip-26-0-169-207:782353:782631 [7] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782631 [7] NCCL INFO Connected all trees +ip-26-0-169-207:782353:782631 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782353:782631 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:782353:782631 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:782351:782650 [5] NCCL INFO comm 0xb2516c0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6c924604558c092c - Init START +ip-26-0-169-207:782351:782650 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:782350:782652 [4] NCCL INFO comm 0xb4af990 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x16618ee55b38bb7e - Init START +ip-26-0-169-207:782350:782652 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782353:782631 [7] NCCL INFO comm 0xb5b5520 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfe3ee127b88530f8 - Init COMPLETE +ip-26-0-169-207:782352:782627 [6] NCCL INFO comm 0xa7934f0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfe3ee127b88530f8 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:782353:782654 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:782349:782633 [3] NCCL INFO Connected all rings +ip-26-0-169-207:782349:782633 [3] NCCL INFO Connected all trees +ip-26-0-169-207:782349:782633 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782349:782633 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:782349:782633 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782353:782654 [7] NCCL INFO comm 0xb5c8b00 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcf42771bb09ccd77 - Init START +ip-26-0-169-207:782352:782656 [6] NCCL INFO comm 0xa7a72c0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8e8262ad0ee14445 - Init START +ip-26-0-169-207:782353:782654 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782352:782656 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782347:782630 [1] NCCL INFO Connected all rings +ip-26-0-169-207:782347:782630 [1] NCCL INFO Connected all trees +ip-26-0-169-207:782347:782630 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782347:782630 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:782347:782630 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782348:782629 [2] NCCL INFO Connected all rings +ip-26-0-169-207:782348:782629 [2] NCCL INFO Connected all trees +ip-26-0-169-207:782348:782629 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782348:782629 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:782348:782629 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782346:782626 [0] NCCL INFO Connected all rings +ip-26-0-169-207:782346:782626 [0] NCCL INFO Connected all trees +ip-26-0-169-207:782346:782626 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782346:782626 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:782346:782626 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782348:782629 [2] NCCL INFO comm 0xb5e9530 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x97a07c9a493552cd - Init COMPLETE +ip-26-0-169-207:782349:782633 [3] NCCL INFO comm 0xa4858b0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x97a07c9a493552cd - Init COMPLETE +ip-26-0-169-207:782347:782630 [1] NCCL INFO comm 0xb14edb0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9d5614c730dd244b - Init COMPLETE +ip-26-0-169-207:782346:782626 [0] NCCL INFO comm 0xb779ff0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9d5614c730dd244b - Init COMPLETE +ip-26-0-169-207:782348:782662 [2] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:782348:782662 [2] NCCL INFO comm 0xb5fd9a0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x17b91ed98339c052 - Init START +ip-26-0-169-207:782348:782662 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:782349:782664 [3] NCCL INFO comm 0xa4991c0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x15d13f2722895787 - Init START +ip-26-0-169-207:782349:782664 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:782347:782666 [1] NCCL INFO comm 0xb1626b0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x53536da8fdddad89 - Init START +ip-26-0-169-207:782347:782666 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:782346:782668 [0] NCCL INFO comm 0xb78dea0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcebe0c0dc024837b - Init START +ip-26-0-169-207:782346:782668 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:782351:782650 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:782351:782650 [5] NCCL INFO Connected all rings +ip-26-0-169-207:782351:782650 [5] NCCL INFO Connected all trees +ip-26-0-169-207:782351:782650 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782350:782652 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:782350:782652 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:782350:782652 [4] NCCL INFO Connected all rings +ip-26-0-169-207:782350:782652 [4] NCCL INFO Connected all trees +ip-26-0-169-207:782350:782652 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782350:782652 [4] NCCL INFO comm 0xb4af990 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x16618ee55b38bb7e - Init COMPLETE +ip-26-0-169-207:782352:782656 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:782352:782656 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:782352:782656 [6] NCCL INFO Connected all rings +ip-26-0-169-207:782352:782656 [6] NCCL INFO Connected all trees +ip-26-0-169-207:782352:782656 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782351:782650 [5] NCCL INFO comm 0xb2516c0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6c924604558c092c - Init COMPLETE +ip-26-0-169-207:782352:782656 [6] NCCL INFO comm 0xa7a72c0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8e8262ad0ee14445 - Init COMPLETE +ip-26-0-169-207:782353:782654 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:782353:782654 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:782353:782654 [7] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782654 [7] NCCL INFO Connected all trees +ip-26-0-169-207:782353:782654 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782353:782654 [7] NCCL INFO comm 0xb5c8b00 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcf42771bb09ccd77 - Init COMPLETE +ip-26-0-169-207:782349:782664 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:782349:782664 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:782349:782664 [3] NCCL INFO Connected all rings +ip-26-0-169-207:782349:782664 [3] NCCL INFO Connected all trees +ip-26-0-169-207:782349:782664 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782348:782662 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:782348:782662 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:782348:782662 [2] NCCL INFO Connected all rings +ip-26-0-169-207:782348:782662 [2] NCCL INFO Connected all trees +ip-26-0-169-207:782348:782662 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782347:782666 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:782347:782666 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:782347:782666 [1] NCCL INFO Connected all rings +ip-26-0-169-207:782347:782666 [1] NCCL INFO Connected all trees +ip-26-0-169-207:782347:782666 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782349:782664 [3] NCCL INFO comm 0xa4991c0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x15d13f2722895787 - Init COMPLETE +ip-26-0-169-207:782346:782668 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:782346:782668 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:782346:782668 [0] NCCL INFO Connected all rings +ip-26-0-169-207:782346:782668 [0] NCCL INFO Connected all trees +ip-26-0-169-207:782346:782668 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782348:782662 [2] NCCL INFO comm 0xb5fd9a0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x17b91ed98339c052 - Init COMPLETE +ip-26-0-169-207:782347:782666 [1] NCCL INFO comm 0xb1626b0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x53536da8fdddad89 - Init COMPLETE +ip-26-0-169-207:782346:782668 [0] NCCL INFO comm 0xb78dea0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcebe0c0dc024837b - Init COMPLETE +12/13/2024 16:39:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Total number of parameters: 1.04G (1984.26MiB) +12/13/2024 16:39:42 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +12/13/2024 16:39:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +12/13/2024 16:39:42 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 15874.00MiB +12/13/2024 16:39:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 17922.00MiB +12/13/2024 16:39:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: No checkpoint path provided. +12/13/2024 16:39:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Parametrizing model parameters using StandardParametrizator +ip-26-0-169-207:782346:782687 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:782352:782688 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:782348:782690 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:782349:782694 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:782353:782692 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:782347:782691 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:782351:782693 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:782350:782695 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:782347:782691 [1] NCCL INFO comm 0xb167df0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7ad7fe5ba661f702 - Init START +ip-26-0-169-207:782349:782694 [3] NCCL INFO comm 0xa49e730 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7ad7fe5ba661f702 - Init START +ip-26-0-169-207:782353:782692 [7] NCCL INFO comm 0xb5ce2d0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7ad7fe5ba661f702 - Init START +ip-26-0-169-207:782351:782693 [5] NCCL INFO comm 0xb256680 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7ad7fe5ba661f702 - Init START +ip-26-0-169-207:782351:782693 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782349:782694 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782353:782692 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782347:782691 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782346:782687 [0] NCCL INFO comm 0xb792ad0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x13370be146797d4b - Init START +ip-26-0-169-207:782348:782690 [2] NCCL INFO comm 0xb602c10 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x13370be146797d4b - Init START +ip-26-0-169-207:782352:782688 [6] NCCL INFO comm 0xa7ac770 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x13370be146797d4b - Init START +ip-26-0-169-207:782350:782695 [4] NCCL INFO comm 0xb4b4ad0 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x13370be146797d4b - Init START +ip-26-0-169-207:782348:782690 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782350:782695 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782352:782688 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:782346:782687 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:782353:782692 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782353:782692 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:782347:782691 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:782352:782688 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782352:782688 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:782346:782687 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:782350:782695 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782350:782695 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:782349:782694 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:782349:782694 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:782351:782693 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:782351:782693 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:782348:782690 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:782348:782690 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:782351:782693 [5] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:782353:782692 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:782351:782693 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:782349:782694 [3] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:782353:782692 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:782349:782694 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:782347:782691 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:782347:782691 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:782348:782690 [2] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:782348:782690 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:782350:782695 [4] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:782352:782688 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:782350:782695 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782352:782688 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:782346:782687 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:782346:782687 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 04/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 05/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 06/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 07/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 04/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 08/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 05/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 09/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 06/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 10/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 07/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 11/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 08/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 12/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 09/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 13/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 10/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 14/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 11/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 15/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 12/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 16/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 13/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 17/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 18/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 14/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 19/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 20/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 04/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 21/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 05/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 15/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 06/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 07/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 22/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 16/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 17/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 18/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 04/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 19/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 05/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 20/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 23/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 08/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 06/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 21/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 09/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 07/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 22/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 10/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 08/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 23/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 11/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 09/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 12/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 10/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 11/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 12/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 13/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 13/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 14/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 15/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 16/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 17/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 18/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 19/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 20/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 14/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 04/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 05/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 06/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 07/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 08/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 09/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 10/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 21/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 15/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 11/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 22/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 16/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 12/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 23/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 17/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 13/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 14/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 18/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 15/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 19/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 20/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 16/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 21/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 22/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 23/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 17/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 18/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782346:782687 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 19/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 20/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 21/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 22/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 23/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Connected all rings +ip-26-0-169-207:782349:782694 [3] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782692 [7] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Connected all rings +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Connected all rings +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 02/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 03/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 04/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 05/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 06/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 07/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 08/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 09/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Connected all rings +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 10/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Connected all rings +ip-26-0-169-207:782346:782687 [0] NCCL INFO Connected all rings +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 11/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 12/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 13/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 14/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 15/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 16/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 17/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 04/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 05/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 18/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 06/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 19/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 20/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 07/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 21/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 08/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 22/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 09/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782352:782688 [6] NCCL INFO Channel 23/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 10/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 11/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 12/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 13/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 14/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 15/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 16/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 17/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 18/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 19/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 20/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 21/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 04/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 05/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 22/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:782351:782693 [5] NCCL INFO Channel 23/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 06/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 07/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 08/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 09/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 10/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 11/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 12/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 13/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 14/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 15/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 16/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 17/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 18/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 19/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 20/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782347:782691 [1] NCCL INFO Connected all trees +ip-26-0-169-207:782347:782691 [1] NCCL INFO NVLS comm 0xb167df0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 21/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 22/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782350:782695 [4] NCCL INFO Channel 23/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782353:782692 [7] NCCL INFO Connected all trees +ip-26-0-169-207:782353:782692 [7] NCCL INFO NVLS comm 0xb5ce2d0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782348:782690 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:782349:782694 [3] NCCL INFO Connected all trees +ip-26-0-169-207:782349:782694 [3] NCCL INFO NVLS comm 0xa49e730 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:782351:782693 [5] NCCL INFO Connected all trees +ip-26-0-169-207:782351:782693 [5] NCCL INFO NVLS comm 0xb256680 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:782352:782688 [6] NCCL INFO Connected all trees +ip-26-0-169-207:782352:782688 [6] NCCL INFO NVLS comm 0xa7ac770 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:782351:782693 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782351:782693 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782351:782693 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:782351:782693 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782349:782694 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782349:782694 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782349:782694 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:782349:782694 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782347:782691 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782347:782691 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782347:782691 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:782347:782691 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782353:782692 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782353:782692 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782353:782692 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:782353:782692 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782349:782694 [3] NCCL INFO comm 0xa49e730 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7ad7fe5ba661f702 - Init COMPLETE +ip-26-0-169-207:782353:782692 [7] NCCL INFO comm 0xb5ce2d0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7ad7fe5ba661f702 - Init COMPLETE +ip-26-0-169-207:782347:782691 [1] NCCL INFO comm 0xb167df0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7ad7fe5ba661f702 - Init COMPLETE +ip-26-0-169-207:782351:782693 [5] NCCL INFO comm 0xb256680 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7ad7fe5ba661f702 - Init COMPLETE +ip-26-0-169-207:782346:782687 [0] NCCL INFO Connected all trees +ip-26-0-169-207:782346:782687 [0] NCCL INFO NVLS comm 0xb792ad0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:782350:782695 [4] NCCL INFO Connected all trees +ip-26-0-169-207:782350:782695 [4] NCCL INFO NVLS comm 0xb4b4ad0 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:782348:782690 [2] NCCL INFO Connected all trees +ip-26-0-169-207:782348:782690 [2] NCCL INFO NVLS comm 0xb602c10 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:782352:782688 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782352:782688 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782352:782688 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:782352:782688 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782348:782690 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782348:782690 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782348:782690 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:782348:782690 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782346:782687 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782346:782687 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782350:782695 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:782350:782695 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:782346:782687 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:782346:782687 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782350:782695 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:782350:782695 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:782348:782690 [2] NCCL INFO comm 0xb602c10 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x13370be146797d4b - Init COMPLETE +ip-26-0-169-207:782346:782687 [0] NCCL INFO comm 0xb792ad0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x13370be146797d4b - Init COMPLETE +ip-26-0-169-207:782352:782688 [6] NCCL INFO comm 0xa7ac770 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x13370be146797d4b - Init COMPLETE +ip-26-0-169-207:782350:782695 [4] NCCL INFO comm 0xb4b4ad0 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x13370be146797d4b - Init COMPLETE +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Optimizer Building] Using LearningRateForSP as learning rate +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] Size of optimizer params per rank: +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 0 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 1 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 2 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 3 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Using dummy data generator +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] There are 1 training stages +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Stage Stable Training Stage] start from step 1 +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: +12/13/2024 16:39:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Start training] datetime: 2024-12-13 16:39:46.706199 | mbs: 2 | grad_accum: 256 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/13/2024 16:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Waiting for 13 seconds +12/13/2024 16:40:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/13/2024 16:40:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3472.48MiB. Peak allocated 5632.00MiB. Peak reserved: 19908.00MiB +12/13/2024 16:40:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3572.67MiB. Peak allocated 6985.16MiB. Peak reserved: 8458.00MiB +12/13/2024 16:40:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 20.8K | tokens_per_sec: 202K | tokens_per_sec_per_gpu: 25.2K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 178 | hardware_tflops_per_gpu: 178 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.91G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 16:40:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.85MiB. Peak allocated 6820.14MiB. Peak reserved: 9452.00MiB +12/13/2024 16:40:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.97MiB. Peak allocated 8000.47MiB. Peak reserved: 9452.00MiB +12/13/2024 16:40:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 18.8K | tokens_per_sec: 223K | tokens_per_sec_per_gpu: 27.9K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.0003 | model_tflops_per_gpu: 197 | hardware_tflops_per_gpu: 197 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 11G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 16:40:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.85MiB. Peak allocated 6820.26MiB. Peak reserved: 10446.00MiB +slurmstepd: error: *** STEP 13417991.0 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T16:40:56 DUE TO TIME LIMIT *** +slurmstepd: error: *** JOB 13417991 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T16:40:56 DUE TO TIME LIMIT *** +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-13 16:40:56,326] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-13 16:40:56,327] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782346 closing signal SIGTERM +[2024-12-13 16:40:56,327] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782347 closing signal SIGTERM +[2024-12-13 16:40:56,327] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782348 closing signal SIGTERM +[2024-12-13 16:40:56,327] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782349 closing signal SIGTERM +[2024-12-13 16:40:56,327] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782350 closing signal SIGTERM +[2024-12-13 16:40:56,327] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782351 closing signal SIGTERM +[2024-12-13 16:40:56,330] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782352 closing signal SIGTERM +[2024-12-13 16:40:56,331] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782353 closing signal SIGTERM +[2024-12-13 16:40:56,486] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782346 closing signal SIGTERM +[2024-12-13 16:40:56,487] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782347 closing signal SIGTERM +[2024-12-13 16:40:56,487] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782348 closing signal SIGTERM +[2024-12-13 16:40:56,487] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782349 closing signal SIGTERM +[2024-12-13 16:40:56,487] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782350 closing signal SIGTERM +[2024-12-13 16:40:56,487] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782351 closing signal SIGTERM +[2024-12-13 16:40:56,487] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782352 closing signal SIGTERM +[2024-12-13 16:40:56,487] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 782353 closing signal SIGTERM diff --git a/logs/13418019-bench_stress_test.out b/logs/13418019-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..a4bfd283e661aae626c61a83b48cf75e6b002bfe --- /dev/null +++ b/logs/13418019-bench_stress_test.out @@ -0,0 +1,3768 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13418019 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 17:14:43,521] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 17:14:43,521] torch.distributed.run: [WARNING] +[2024-12-13 17:14:43,521] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 17:14:43,521] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 17:14:43,521] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-169-207:797061:797061 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:797061:797061 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:797061:797061 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:797061:797061 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:797061:797061 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:797065:797065 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:797062:797062 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:797065:797065 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:797062:797062 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:797064:797064 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:797067:797067 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:797064:797064 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:797067:797067 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:797068:797068 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:797068:797068 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:797063:797063 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:797063:797063 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:797066:797066 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:797065:797065 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:797062:797062 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:797066:797066 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:797062:797062 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:797065:797065 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:797065:797065 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:797062:797062 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:797064:797064 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:797067:797067 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:797068:797068 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:797064:797064 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:797064:797064 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:797067:797067 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:797067:797067 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:797063:797063 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:797066:797066 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:797068:797068 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:797068:797068 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:797066:797066 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:797063:797063 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:797066:797066 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:797063:797063 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:797062:797184 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:797062:797184 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/162 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:797063:797187 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:797063:797187 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:797061:797182 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:797061:797182 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:797065:797183 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:797065:797183 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:797066:797186 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:797066:797186 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:797067:797188 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:797067:797188 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:797068:797189 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:797068:797189 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:797064:797185 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:797064:797185 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:797068:797189 [7] NCCL INFO comm 0x98ac590 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa0c2660fb46d2158 - Init START +ip-26-0-169-207:797061:797182 [0] NCCL INFO comm 0x8f1fb50 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa0c2660fb46d2158 - Init START +ip-26-0-169-207:797063:797187 [2] NCCL INFO comm 0x98ebe70 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa0c2660fb46d2158 - Init START +ip-26-0-169-207:797064:797185 [3] NCCL INFO comm 0x89bb120 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa0c2660fb46d2158 - Init START +ip-26-0-169-207:797067:797188 [6] NCCL INFO comm 0x8890a80 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa0c2660fb46d2158 - Init START +ip-26-0-169-207:797065:797183 [4] NCCL INFO comm 0x897eee0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa0c2660fb46d2158 - Init START +ip-26-0-169-207:797066:797186 [5] NCCL INFO comm 0x93aa8a0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa0c2660fb46d2158 - Init START +ip-26-0-169-207:797062:797184 [1] NCCL INFO comm 0x9d13610 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa0c2660fb46d2158 - Init START +ip-26-0-169-207:797062:797184 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797063:797187 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797061:797182 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797064:797185 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797065:797183 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797065:797183 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:797067:797188 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:797066:797186 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797066:797186 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:797068:797189 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797068:797189 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:797062:797184 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:797062:797184 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:797061:797182 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:797063:797187 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:797063:797187 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:797064:797185 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:797064:797185 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:797065:797183 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:797064:797185 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:797066:797186 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:797065:797183 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797064:797185 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797067:797188 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:797066:797186 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797067:797188 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797068:797189 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797068:797189 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797063:797187 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797063:797187 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797062:797184 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:797061:797182 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:797061:797182 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797062:797184 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Connected all rings +ip-26-0-169-207:797061:797182 [0] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797189 [7] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Connected all rings +ip-26-0-169-207:797067:797188 [6] NCCL INFO Connected all rings +ip-26-0-169-207:797066:797186 [5] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Connected all rings +ip-26-0-169-207:797064:797185 [3] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797188 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797187 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797064:797185 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797183 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797061:797182 [0] NCCL INFO Connected all trees +ip-26-0-169-207:797061:797182 [0] NCCL INFO NVLS comm 0x8f1fb50 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO Connected all trees +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797068:797189 [7] NCCL INFO NVLS comm 0x98ac590 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797186 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797184 [1] NCCL INFO Connected all trees +ip-26-0-169-207:797062:797184 [1] NCCL INFO NVLS comm 0x9d13610 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797064:797185 [3] NCCL INFO Connected all trees +ip-26-0-169-207:797064:797185 [3] NCCL INFO NVLS comm 0x89bb120 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797063:797187 [2] NCCL INFO Connected all trees +ip-26-0-169-207:797063:797187 [2] NCCL INFO NVLS comm 0x98ebe70 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797067:797188 [6] NCCL INFO Connected all trees +ip-26-0-169-207:797067:797188 [6] NCCL INFO NVLS comm 0x8890a80 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797066:797186 [5] NCCL INFO Connected all trees +ip-26-0-169-207:797066:797186 [5] NCCL INFO NVLS comm 0x93aa8a0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797065:797183 [4] NCCL INFO Connected all trees +ip-26-0-169-207:797065:797183 [4] NCCL INFO NVLS comm 0x897eee0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797068:797189 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797068:797189 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797068:797189 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797068:797189 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797062:797184 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797065:797183 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797066:797186 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797065:797183 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797065:797183 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797065:797183 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797062:797184 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797061:797182 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797066:797186 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797061:797182 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797066:797186 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797066:797186 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797062:797184 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797063:797187 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797062:797184 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797067:797188 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797063:797187 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797063:797187 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797067:797188 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797063:797187 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797067:797188 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797067:797188 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797061:797182 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797061:797182 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797064:797185 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797064:797185 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797064:797185 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797064:797185 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797064:797185 [3] NCCL INFO comm 0x89bb120 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa0c2660fb46d2158 - Init COMPLETE +ip-26-0-169-207:797068:797189 [7] NCCL INFO comm 0x98ac590 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa0c2660fb46d2158 - Init COMPLETE +ip-26-0-169-207:797063:797187 [2] NCCL INFO comm 0x98ebe70 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa0c2660fb46d2158 - Init COMPLETE +ip-26-0-169-207:797067:797188 [6] NCCL INFO comm 0x8890a80 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa0c2660fb46d2158 - Init COMPLETE +ip-26-0-169-207:797066:797186 [5] NCCL INFO comm 0x93aa8a0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa0c2660fb46d2158 - Init COMPLETE +ip-26-0-169-207:797065:797183 [4] NCCL INFO comm 0x897eee0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa0c2660fb46d2158 - Init COMPLETE +ip-26-0-169-207:797061:797182 [0] NCCL INFO comm 0x8f1fb50 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa0c2660fb46d2158 - Init COMPLETE +ip-26-0-169-207:797062:797184 [1] NCCL INFO comm 0x9d13610 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa0c2660fb46d2158 - Init COMPLETE +ip-26-0-169-207:797061:797257 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:797062:797261 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:797067:797258 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:797065:797263 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:797064:797259 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:797068:797260 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:797063:797262 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:797066:797264 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:797062:797261 [1] NCCL INFO comm 0x9f3d4c0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6454054e1e0ff82f - Init START +ip-26-0-169-207:797063:797262 [2] NCCL INFO comm 0x9b15de0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6454054e1e0ff82f - Init START +ip-26-0-169-207:797064:797259 [3] NCCL INFO comm 0x8be4e80 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6454054e1e0ff82f - Init START +ip-26-0-169-207:797065:797263 [4] NCCL INFO comm 0x8ba8a80 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6454054e1e0ff82f - Init START +ip-26-0-169-207:797068:797260 [7] NCCL INFO comm 0x9ad6180 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6454054e1e0ff82f - Init START +ip-26-0-169-207:797061:797257 [0] NCCL INFO comm 0x9149690 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6454054e1e0ff82f - Init START +ip-26-0-169-207:797066:797264 [5] NCCL INFO comm 0x95d4cf0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6454054e1e0ff82f - Init START +ip-26-0-169-207:797067:797258 [6] NCCL INFO comm 0x8aba6d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6454054e1e0ff82f - Init START +ip-26-0-169-207:797065:797263 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797064:797259 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797061:797257 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:797062:797261 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797067:797258 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797068:797260 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797063:797262 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797066:797264 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797067:797258 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797067:797258 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:797066:797264 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797066:797264 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:797063:797262 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:797063:797262 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:797061:797257 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:797064:797259 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:797064:797259 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:797068:797260 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797068:797260 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:797065:797263 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797065:797263 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:797062:797261 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:797062:797261 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:797062:797261 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:797062:797261 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797063:797262 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:797063:797262 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797068:797260 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797068:797260 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797067:797258 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797067:797258 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797066:797264 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797065:797263 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797066:797264 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797065:797263 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797257 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:797064:797259 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:797061:797257 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797064:797259 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Connected all rings +ip-26-0-169-207:797064:797259 [3] NCCL INFO Connected all rings +ip-26-0-169-207:797062:797261 [1] NCCL INFO Connected all rings +ip-26-0-169-207:797063:797262 [2] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797260 [7] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797068:797260 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797263 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797063:797262 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797261 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797259 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797258 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797264 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797061:797257 [0] NCCL INFO Connected all trees +ip-26-0-169-207:797061:797257 [0] NCCL INFO NVLS comm 0x9149690 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797068:797260 [7] NCCL INFO Connected all trees +ip-26-0-169-207:797068:797260 [7] NCCL INFO NVLS comm 0x9ad6180 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797063:797262 [2] NCCL INFO Connected all trees +ip-26-0-169-207:797063:797262 [2] NCCL INFO NVLS comm 0x9b15de0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797064:797259 [3] NCCL INFO Connected all trees +ip-26-0-169-207:797064:797259 [3] NCCL INFO NVLS comm 0x8be4e80 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797062:797261 [1] NCCL INFO Connected all trees +ip-26-0-169-207:797062:797261 [1] NCCL INFO NVLS comm 0x9f3d4c0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797066:797264 [5] NCCL INFO Connected all trees +ip-26-0-169-207:797066:797264 [5] NCCL INFO NVLS comm 0x95d4cf0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797067:797258 [6] NCCL INFO Connected all trees +ip-26-0-169-207:797067:797258 [6] NCCL INFO NVLS comm 0x8aba6d0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797065:797263 [4] NCCL INFO Connected all trees +ip-26-0-169-207:797065:797263 [4] NCCL INFO NVLS comm 0x8ba8a80 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797065:797263 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797065:797263 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797065:797263 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797065:797263 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797062:797261 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797062:797261 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797062:797261 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797062:797261 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797063:797262 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797063:797262 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797068:797260 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797063:797262 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797063:797262 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797066:797264 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797068:797260 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797068:797260 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797068:797260 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797066:797264 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797067:797258 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797067:797258 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797066:797264 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797066:797264 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797067:797258 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797067:797258 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797064:797259 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797064:797259 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797064:797259 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797064:797259 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797061:797257 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797061:797257 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797061:797257 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797061:797257 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797068:797260 [7] NCCL INFO comm 0x9ad6180 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6454054e1e0ff82f - Init COMPLETE +ip-26-0-169-207:797066:797264 [5] NCCL INFO comm 0x95d4cf0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6454054e1e0ff82f - Init COMPLETE +ip-26-0-169-207:797067:797258 [6] NCCL INFO comm 0x8aba6d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6454054e1e0ff82f - Init COMPLETE +ip-26-0-169-207:797064:797259 [3] NCCL INFO comm 0x8be4e80 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6454054e1e0ff82f - Init COMPLETE +ip-26-0-169-207:797065:797263 [4] NCCL INFO comm 0x8ba8a80 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6454054e1e0ff82f - Init COMPLETE +ip-26-0-169-207:797063:797262 [2] NCCL INFO comm 0x9b15de0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6454054e1e0ff82f - Init COMPLETE +ip-26-0-169-207:797062:797261 [1] NCCL INFO comm 0x9f3d4c0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6454054e1e0ff82f - Init COMPLETE +ip-26-0-169-207:797061:797257 [0] NCCL INFO comm 0x9149690 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6454054e1e0ff82f - Init COMPLETE +ip-26-0-169-207:797061:797290 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:797062:797294 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:797066:797297 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:797067:797293 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:797065:797292 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:797064:797295 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:797063:797296 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:797068:797291 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:797068:797291 [7] NCCL INFO comm 0x9aea680 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe7f8e10f74710eb9 - Init START +ip-26-0-169-207:797065:797292 [4] NCCL INFO comm 0x8bbd210 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe7f8e10f74710eb9 - Init START +ip-26-0-169-207:797067:797293 [6] NCCL INFO comm 0x8acee30 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe7f8e10f74710eb9 - Init START +ip-26-0-169-207:797061:797290 [0] NCCL INFO comm 0x915dd00 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe7f8e10f74710eb9 - Init START +ip-26-0-169-207:797066:797297 [5] NCCL INFO comm 0x95e90d0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe7f8e10f74710eb9 - Init START +ip-26-0-169-207:797063:797296 [2] NCCL INFO comm 0x9b2a420 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe7f8e10f74710eb9 - Init START +ip-26-0-169-207:797062:797294 [1] NCCL INFO comm 0x9f51a50 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe7f8e10f74710eb9 - Init START +ip-26-0-169-207:797064:797295 [3] NCCL INFO comm 0x8bf9780 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe7f8e10f74710eb9 - Init START +ip-26-0-169-207:797068:797291 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797065:797292 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797063:797296 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797062:797294 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797064:797295 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797066:797297 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797067:797293 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797061:797290 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:797063:797296 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:797063:797296 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:797066:797297 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:797066:797297 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:797061:797290 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:797067:797293 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797067:797293 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:797062:797294 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:797062:797294 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:797068:797291 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797068:797291 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:797064:797295 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:797064:797295 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:797065:797292 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797065:797292 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:797065:797292 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:797065:797292 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797066:797297 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:797066:797297 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797067:797293 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:797067:797293 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797063:797296 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:797063:797296 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797068:797291 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797068:797291 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797062:797294 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:797064:797295 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797062:797294 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797064:797295 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:797061:797290 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Connected all rings +ip-26-0-169-207:797063:797296 [2] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797291 [7] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Connected all rings +ip-26-0-169-207:797065:797292 [4] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Connected all rings +ip-26-0-169-207:797066:797297 [5] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797061:797290 [0] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797063:797296 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797292 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797067:797293 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797062:797294 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797066:797297 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797064:797295 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:797068:797291 [7] NCCL INFO Connected all trees +ip-26-0-169-207:797068:797291 [7] NCCL INFO NVLS comm 0x9aea680 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797061:797290 [0] NCCL INFO Connected all trees +ip-26-0-169-207:797061:797290 [0] NCCL INFO NVLS comm 0x915dd00 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797063:797296 [2] NCCL INFO Connected all trees +ip-26-0-169-207:797063:797296 [2] NCCL INFO NVLS comm 0x9b2a420 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797062:797294 [1] NCCL INFO Connected all trees +ip-26-0-169-207:797062:797294 [1] NCCL INFO NVLS comm 0x9f51a50 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797067:797293 [6] NCCL INFO Connected all trees +ip-26-0-169-207:797067:797293 [6] NCCL INFO NVLS comm 0x8acee30 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797065:797292 [4] NCCL INFO Connected all trees +ip-26-0-169-207:797065:797292 [4] NCCL INFO NVLS comm 0x8bbd210 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797064:797295 [3] NCCL INFO Connected all trees +ip-26-0-169-207:797064:797295 [3] NCCL INFO NVLS comm 0x8bf9780 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797066:797297 [5] NCCL INFO Connected all trees +ip-26-0-169-207:797066:797297 [5] NCCL INFO NVLS comm 0x95e90d0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:797068:797291 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797068:797291 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797068:797291 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797068:797291 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797062:797294 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797064:797295 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797062:797294 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797064:797295 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797062:797294 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797062:797294 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797064:797295 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797064:797295 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797066:797297 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797066:797297 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797066:797297 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797066:797297 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797065:797292 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797065:797292 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797065:797292 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797065:797292 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797063:797296 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797063:797296 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797063:797296 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797063:797296 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797067:797293 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797067:797293 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797067:797293 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797067:797293 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797061:797290 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797061:797290 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797061:797290 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:797061:797290 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797063:797296 [2] NCCL INFO comm 0x9b2a420 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe7f8e10f74710eb9 - Init COMPLETE +ip-26-0-169-207:797067:797293 [6] NCCL INFO comm 0x8acee30 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe7f8e10f74710eb9 - Init COMPLETE +ip-26-0-169-207:797061:797290 [0] NCCL INFO comm 0x915dd00 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe7f8e10f74710eb9 - Init COMPLETE +ip-26-0-169-207:797065:797292 [4] NCCL INFO comm 0x8bbd210 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe7f8e10f74710eb9 - Init COMPLETE +ip-26-0-169-207:797064:797295 [3] NCCL INFO comm 0x8bf9780 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe7f8e10f74710eb9 - Init COMPLETE +ip-26-0-169-207:797068:797291 [7] NCCL INFO comm 0x9aea680 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe7f8e10f74710eb9 - Init COMPLETE +ip-26-0-169-207:797066:797297 [5] NCCL INFO comm 0x95e90d0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe7f8e10f74710eb9 - Init COMPLETE +ip-26-0-169-207:797062:797294 [1] NCCL INFO comm 0x9f51a50 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe7f8e10f74710eb9 - Init COMPLETE +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config: +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config(general=GeneralArgs(project='debug', +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: run='1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k', +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: step=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: consumed_train_samples=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ignore_sanity_checks=True), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: parallelism=ParallelismArgs(dp=4, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp=1, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp=2, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp_engine=, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_mode=, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_linear_async_communication=True, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: recompute_layer=False, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_recompute_allgather=True, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: expert_parallel_size=1), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: init_method=RandomInit(std=0.02), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: dtype=torch.bfloat16, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: make_vocab_size_divisible_by=1, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ddp_bucket_cap_mb=25), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_revision=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_max_length=None), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoint_interval=10000, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_initial_state=False, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_final_state=False, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: resume_checkpoint_path=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints_path_is_shared_file_system=False), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: logging=LoggingArgs(log_level='info', +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: log_level_replica='info', +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration_step_info_interval=1), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokens=TokensArgs(sequence_length=2048, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: train_steps=100, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: micro_batch_size=2, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: batch_accumulation_per_replica=256, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: val_check_interval=100, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_val_batches=0, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_test_batches=0), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta1=0.9, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta2=0.95, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: torch_adam_is_fused=True, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: name='adamW'), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: zero_stage=1, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: weight_decay=0.01, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: clip_grad=1.0, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: accumulate_grad_in_fp32=True, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_steps=2, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_style='linear', +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_style='cosine', +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_steps=13, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_starting_step=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: min_decay_lr=1e-05)), +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: start_training_step=1, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data=DataArgs(dataset=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_loading_workers=1))], +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: profiler=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lighteval=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: s3_upload=None) +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Model Config: +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: LlamaConfig(bos_token_id=0, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768) +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Building model.. +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Initialize RoPE Theta = 10000.0 +12/13/2024 17:15:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:797067:797345 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:797062:797349 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:797068:797348 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:797064:797350 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:797066:797351 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:797061:797347 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:797063:797346 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:797067:797345 [6] NCCL INFO comm 0xa8d0cb0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd1856dc9ae550331 - Init START +ip-26-0-169-207:797068:797348 [7] NCCL INFO comm 0xb8ec6a0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd1856dc9ae550331 - Init START +ip-26-0-169-207:797067:797345 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797065:797344 [4] NCCL INFO comm 0xa9bfb20 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe7c6dcbf22cf7179 - Init START +ip-26-0-169-207:797066:797351 [5] NCCL INFO comm 0xb3e9470 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe7c6dcbf22cf7179 - Init START +ip-26-0-169-207:797061:797347 [0] NCCL INFO comm 0xaf60d30 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2ee4375d26d98b90 - Init START +ip-26-0-169-207:797062:797349 [1] NCCL INFO comm 0xbd55760 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2ee4375d26d98b90 - Init START +ip-26-0-169-207:797065:797344 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797061:797347 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:797063:797346 [2] NCCL INFO comm 0xb92b3b0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe6bee6bfb59477e3 - Init START +ip-26-0-169-207:797064:797350 [3] NCCL INFO comm 0xa9fb130 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe6bee6bfb59477e3 - Init START +ip-26-0-169-207:797063:797346 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797068:797348 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797062:797349 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797066:797351 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797064:797350 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797062:797349 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:797063:797346 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:797067:797345 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797066:797351 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797064:797350 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:797064:797350 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:797064:797350 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:797063:797346 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:797062:797349 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:797062:797349 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:797061:797347 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:797066:797351 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:797066:797351 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797065:797344 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:797065:797344 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797068:797348 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:797068:797348 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:797068:797348 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:797067:797345 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797064:797350 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797347 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797062:797349 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797065:797344 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797068:797348 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797066:797351 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:797067:797345 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:797063:797346 [2] NCCL INFO Connected all rings +ip-26-0-169-207:797063:797346 [2] NCCL INFO Connected all trees +ip-26-0-169-207:797063:797346 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797063:797346 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:797063:797346 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797061:797347 [0] NCCL INFO Connected all rings +ip-26-0-169-207:797061:797347 [0] NCCL INFO Connected all trees +ip-26-0-169-207:797061:797347 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797061:797347 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:797061:797347 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797062:797349 [1] NCCL INFO Connected all rings +ip-26-0-169-207:797062:797349 [1] NCCL INFO Connected all trees +ip-26-0-169-207:797062:797349 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797062:797349 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:797062:797349 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797068:797348 [7] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797348 [7] NCCL INFO Connected all trees +ip-26-0-169-207:797068:797348 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797068:797348 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:797068:797348 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797062:797349 [1] NCCL INFO comm 0xbd55760 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2ee4375d26d98b90 - Init COMPLETE +ip-26-0-169-207:797061:797347 [0] NCCL INFO comm 0xaf60d30 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2ee4375d26d98b90 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:797061:797368 [0] NCCL INFO comm 0xaf75290 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3e5799f50c227508 - Init START +ip-26-0-169-207:797061:797368 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:797062:797370 [1] NCCL INFO comm 0xbd69050 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfb932c8818014040 - Init START +ip-26-0-169-207:797062:797370 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797064:797350 [3] NCCL INFO Connected all rings +ip-26-0-169-207:797064:797350 [3] NCCL INFO Connected all trees +ip-26-0-169-207:797064:797350 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797064:797350 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:797064:797350 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797067:797345 [6] NCCL INFO Connected all rings +ip-26-0-169-207:797067:797345 [6] NCCL INFO Connected all trees +ip-26-0-169-207:797067:797345 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797067:797345 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:797067:797345 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797065:797344 [4] NCCL INFO Connected all rings +ip-26-0-169-207:797065:797344 [4] NCCL INFO Connected all trees +ip-26-0-169-207:797065:797344 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797065:797344 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:797065:797344 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797064:797350 [3] NCCL INFO comm 0xa9fb130 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe6bee6bfb59477e3 - Init COMPLETE +ip-26-0-169-207:797063:797346 [2] NCCL INFO comm 0xb92b3b0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe6bee6bfb59477e3 - Init COMPLETE +ip-26-0-169-207:797066:797351 [5] NCCL INFO Connected all rings +ip-26-0-169-207:797066:797351 [5] NCCL INFO Connected all trees +ip-26-0-169-207:797066:797351 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797066:797351 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:797066:797351 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797063:797376 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:797063:797376 [2] NCCL INFO comm 0xb93f720 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc270c47d6600b3b6 - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:797063:797376 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:797064:797378 [3] NCCL INFO comm 0xaa0eae0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd794c42d12db5aa4 - Init START +ip-26-0-169-207:797064:797378 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797067:797345 [6] NCCL INFO comm 0xa8d0cb0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd1856dc9ae550331 - Init COMPLETE +ip-26-0-169-207:797068:797348 [7] NCCL INFO comm 0xb8ec6a0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd1856dc9ae550331 - Init COMPLETE +ip-26-0-169-207:797065:797344 [4] NCCL INFO comm 0xa9bfb20 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe7c6dcbf22cf7179 - Init COMPLETE +ip-26-0-169-207:797066:797351 [5] NCCL INFO comm 0xb3e9470 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe7c6dcbf22cf7179 - Init COMPLETE +ip-26-0-169-207:797065:797380 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:797065:797380 [4] NCCL INFO comm 0xa9d3d70 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x15df45453266d751 - Init START +ip-26-0-169-207:797065:797380 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:797066:797382 [5] NCCL INFO comm 0xb3fce60 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfe4ee068be0ee6f - Init START +ip-26-0-169-207:797066:797382 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:797067:797385 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:797068:797386 [7] NCCL INFO comm 0xb8fff60 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2f591e12139efd57 - Init START +ip-26-0-169-207:797068:797386 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797067:797385 [6] NCCL INFO comm 0xa8e5020 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x327df9b163322641 - Init START +ip-26-0-169-207:797067:797385 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:797061:797368 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:797061:797368 [0] NCCL INFO Connected all rings +ip-26-0-169-207:797061:797368 [0] NCCL INFO Connected all trees +ip-26-0-169-207:797061:797368 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797062:797370 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:797062:797370 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:797062:797370 [1] NCCL INFO Connected all rings +ip-26-0-169-207:797062:797370 [1] NCCL INFO Connected all trees +ip-26-0-169-207:797062:797370 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797061:797368 [0] NCCL INFO comm 0xaf75290 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3e5799f50c227508 - Init COMPLETE +ip-26-0-169-207:797062:797370 [1] NCCL INFO comm 0xbd69050 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfb932c8818014040 - Init COMPLETE +ip-26-0-169-207:797064:797378 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:797064:797378 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:797064:797378 [3] NCCL INFO P2P Chunksize set to 131072 +12/13/2024 17:15:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Total number of parameters: 1.04G (1984.26MiB) +12/13/2024 17:15:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +12/13/2024 17:15:21 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +ip-26-0-169-207:797064:797378 [3] NCCL INFO Connected all rings +ip-26-0-169-207:797064:797378 [3] NCCL INFO Connected all trees +ip-26-0-169-207:797064:797378 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +12/13/2024 17:15:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/13/2024 17:15:21 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 17922.00MiB +12/13/2024 17:15:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: No checkpoint path provided. +12/13/2024 17:15:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Parametrizing model parameters using StandardParametrizator +ip-26-0-169-207:797066:797382 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:797066:797382 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:797066:797382 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:797066:797382 [5] NCCL INFO Connected all rings +ip-26-0-169-207:797066:797382 [5] NCCL INFO Connected all trees +ip-26-0-169-207:797066:797382 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797062:797398 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:797067:797385 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:797067:797385 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:797067:797385 [6] NCCL INFO Connected all rings +ip-26-0-169-207:797067:797385 [6] NCCL INFO Connected all trees +ip-26-0-169-207:797067:797385 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797063:797376 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:797063:797376 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:797063:797376 [2] NCCL INFO Connected all rings +ip-26-0-169-207:797063:797376 [2] NCCL INFO Connected all trees +ip-26-0-169-207:797063:797376 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797064:797378 [3] NCCL INFO comm 0xaa0eae0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd794c42d12db5aa4 - Init COMPLETE +ip-26-0-169-207:797065:797380 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:797065:797380 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:797065:797380 [4] NCCL INFO Connected all rings +ip-26-0-169-207:797065:797380 [4] NCCL INFO Connected all trees +ip-26-0-169-207:797065:797380 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797068:797386 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:797068:797386 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:797068:797386 [7] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797386 [7] NCCL INFO Connected all trees +ip-26-0-169-207:797068:797386 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797066:797382 [5] NCCL INFO comm 0xb3fce60 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfe4ee068be0ee6f - Init COMPLETE +ip-26-0-169-207:797067:797385 [6] NCCL INFO comm 0xa8e5020 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x327df9b163322641 - Init COMPLETE +ip-26-0-169-207:797063:797376 [2] NCCL INFO comm 0xb93f720 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc270c47d6600b3b6 - Init COMPLETE +ip-26-0-169-207:797065:797380 [4] NCCL INFO comm 0xa9d3d70 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x15df45453266d751 - Init COMPLETE +ip-26-0-169-207:797068:797386 [7] NCCL INFO comm 0xb8fff60 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2f591e12139efd57 - Init COMPLETE +ip-26-0-169-207:797063:797408 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:797064:797409 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:797065:797410 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:797066:797411 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:797068:797413 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:797067:797412 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:797068:797413 [7] NCCL INFO comm 0xb904c30 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x48ff116c1e30855 - Init START +ip-26-0-169-207:797064:797409 [3] NCCL INFO comm 0xaa141e0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x48ff116c1e30855 - Init START +ip-26-0-169-207:797066:797411 [5] NCCL INFO comm 0xb401f60 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x48ff116c1e30855 - Init START +ip-26-0-169-207:797062:797398 [1] NCCL INFO comm 0xbd6e290 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x48ff116c1e30855 - Init START +ip-26-0-169-207:797068:797413 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797064:797409 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797066:797411 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797062:797398 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797067:797412 [6] NCCL INFO comm 0xa8e97e0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe16cfb510ef45e09 - Init START +ip-26-0-169-207:797065:797410 [4] NCCL INFO comm 0xa9d88f0 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe16cfb510ef45e09 - Init START +ip-26-0-169-207:797063:797408 [2] NCCL INFO comm 0xb944060 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe16cfb510ef45e09 - Init START +ip-26-0-169-207:797061:797395 [0] NCCL INFO comm 0xaf7a760 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe16cfb510ef45e09 - Init START +ip-26-0-169-207:797067:797412 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797063:797408 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797065:797410 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:797061:797395 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:797068:797413 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797068:797413 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:797067:797412 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797067:797412 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:797065:797410 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797065:797410 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:797064:797409 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:797064:797409 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:797063:797408 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:797063:797408 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:797061:797395 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:797062:797398 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:797066:797411 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:797066:797411 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:797063:797408 [2] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:797067:797412 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:797065:797410 [4] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:797067:797412 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797063:797408 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797065:797410 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:797061:797395 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:797068:797413 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:797066:797411 [5] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:797068:797413 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:797064:797409 [3] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:797066:797411 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:797064:797409 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:797062:797398 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 04/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 05/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 06/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 07/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 08/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 04/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 09/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 05/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 10/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 06/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 11/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 07/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 12/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 08/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 13/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 04/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 09/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 14/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 05/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 10/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 15/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 06/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 11/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 16/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 07/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 17/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 18/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 12/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 13/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 14/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 15/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 08/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 16/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 19/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 04/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 17/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 09/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 05/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 20/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 18/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 10/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 06/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 21/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 19/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 11/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 22/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 07/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 20/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 12/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 23/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 08/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 13/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 21/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 09/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 22/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 14/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 10/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 04/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 23/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 15/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 11/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 05/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 16/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 12/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 06/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 17/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 13/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 07/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 18/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 19/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 20/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 14/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 08/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 21/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 15/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 09/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 22/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 16/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 10/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 23/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 17/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 11/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 12/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 18/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 13/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 19/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 14/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 20/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 15/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 21/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 16/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 22/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 17/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 18/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 23/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 19/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 20/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 21/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 22/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 23/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Connected all rings +ip-26-0-169-207:797065:797410 [4] NCCL INFO Connected all rings +ip-26-0-169-207:797061:797395 [0] NCCL INFO Connected all rings +ip-26-0-169-207:797067:797412 [6] NCCL INFO Connected all rings +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 02/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 03/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 04/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 05/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 06/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 07/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 08/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 09/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 10/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 11/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Connected all rings +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 12/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 13/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797062:797398 [1] NCCL INFO Connected all rings +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 14/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 15/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 16/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 17/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 18/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 19/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 20/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 21/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 22/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797067:797412 [6] NCCL INFO Channel 23/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 04/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 05/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 06/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 07/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 08/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 09/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 10/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 11/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 12/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 13/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797068:797413 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 14/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 15/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 04/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 16/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 05/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 17/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 06/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 18/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797063:797408 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 07/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 19/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 08/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 20/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 09/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 21/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 10/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 22/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 11/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797065:797410 [4] NCCL INFO Channel 23/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 12/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 13/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 14/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 15/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 16/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 17/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 18/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 19/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 20/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 21/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797064:797409 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 22/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797066:797411 [5] NCCL INFO Channel 23/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:797061:797395 [0] NCCL INFO Connected all trees +ip-26-0-169-207:797061:797395 [0] NCCL INFO NVLS comm 0xaf7a760 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:797067:797412 [6] NCCL INFO Connected all trees +ip-26-0-169-207:797067:797412 [6] NCCL INFO NVLS comm 0xa8e97e0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:797068:797413 [7] NCCL INFO Connected all trees +ip-26-0-169-207:797068:797413 [7] NCCL INFO NVLS comm 0xb904c30 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:797063:797408 [2] NCCL INFO Connected all trees +ip-26-0-169-207:797063:797408 [2] NCCL INFO NVLS comm 0xb944060 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:797062:797398 [1] NCCL INFO Connected all trees +ip-26-0-169-207:797062:797398 [1] NCCL INFO NVLS comm 0xbd6e290 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:797065:797410 [4] NCCL INFO Connected all trees +ip-26-0-169-207:797065:797410 [4] NCCL INFO NVLS comm 0xa9d88f0 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:797067:797412 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797067:797412 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797067:797412 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:797067:797412 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797063:797408 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797063:797408 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797063:797408 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:797063:797408 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797065:797410 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797065:797410 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797065:797410 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:797065:797410 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797061:797395 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797061:797395 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797061:797395 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:797061:797395 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797064:797409 [3] NCCL INFO Connected all trees +ip-26-0-169-207:797064:797409 [3] NCCL INFO NVLS comm 0xaa141e0 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:797066:797411 [5] NCCL INFO Connected all trees +ip-26-0-169-207:797066:797411 [5] NCCL INFO NVLS comm 0xb401f60 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:797063:797408 [2] NCCL INFO comm 0xb944060 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe16cfb510ef45e09 - Init COMPLETE +ip-26-0-169-207:797061:797395 [0] NCCL INFO comm 0xaf7a760 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe16cfb510ef45e09 - Init COMPLETE +ip-26-0-169-207:797067:797412 [6] NCCL INFO comm 0xa8e97e0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe16cfb510ef45e09 - Init COMPLETE +ip-26-0-169-207:797065:797410 [4] NCCL INFO comm 0xa9d88f0 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe16cfb510ef45e09 - Init COMPLETE +ip-26-0-169-207:797068:797413 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797068:797413 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797066:797411 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797068:797413 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:797068:797413 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797066:797411 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797066:797411 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:797066:797411 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797064:797409 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797062:797398 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:797064:797409 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797062:797398 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:797064:797409 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:797064:797409 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797062:797398 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:797062:797398 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:797068:797413 [7] NCCL INFO comm 0xb904c30 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x48ff116c1e30855 - Init COMPLETE +ip-26-0-169-207:797062:797398 [1] NCCL INFO comm 0xbd6e290 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x48ff116c1e30855 - Init COMPLETE +ip-26-0-169-207:797064:797409 [3] NCCL INFO comm 0xaa141e0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x48ff116c1e30855 - Init COMPLETE +ip-26-0-169-207:797066:797411 [5] NCCL INFO comm 0xb401f60 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x48ff116c1e30855 - Init COMPLETE +12/13/2024 17:15:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Optimizer Building] Using LearningRateForSP as learning rate +12/13/2024 17:15:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] Size of optimizer params per rank: +12/13/2024 17:15:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 0 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 17:15:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 1 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 17:15:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 2 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 17:15:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 3 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 17:15:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/13/2024 17:15:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Using dummy data generator +12/13/2024 17:15:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] There are 1 training stages +12/13/2024 17:15:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Stage Stable Training Stage] start from step 1 +12/13/2024 17:15:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: +12/13/2024 17:15:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Start training] datetime: 2024-12-13 17:15:26.018656 | mbs: 2 | grad_accum: 256 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/13/2024 17:15:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Waiting for 16 seconds +12/13/2024 17:15:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/13/2024 17:15:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3472.48MiB. Peak allocated 5632.00MiB. Peak reserved: 18884.00MiB +12/13/2024 17:16:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3572.67MiB. Peak allocated 6985.16MiB. Peak reserved: 7434.00MiB +12/13/2024 17:16:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 20.6K | tokens_per_sec: 203K | tokens_per_sec_per_gpu: 25.4K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 179 | hardware_tflops_per_gpu: 179 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.62G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 17:16:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.85MiB. Peak allocated 6820.14MiB. Peak reserved: 9178.00MiB +12/13/2024 17:16:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.97MiB. Peak allocated 8000.47MiB. Peak reserved: 9178.00MiB +12/13/2024 17:16:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 18.4K | tokens_per_sec: 228K | tokens_per_sec_per_gpu: 28.5K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.0003 | model_tflops_per_gpu: 201 | hardware_tflops_per_gpu: 201 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.62G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 17:16:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.85MiB. Peak allocated 6820.26MiB. Peak reserved: 9178.00MiB +slurmstepd: error: *** JOB 13418019 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T17:16:41 DUE TO TIME LIMIT *** +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +slurmstepd: error: *** STEP 13418019.0 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T17:16:41 DUE TO TIME LIMIT *** +[2024-12-13 17:16:41,957] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-13 17:16:41,957] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797061 closing signal SIGTERM +[2024-12-13 17:16:41,957] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797062 closing signal SIGTERM +[2024-12-13 17:16:41,957] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797063 closing signal SIGTERM +[2024-12-13 17:16:41,957] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797064 closing signal SIGTERM +[2024-12-13 17:16:41,957] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797065 closing signal SIGTERM +[2024-12-13 17:16:41,958] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797066 closing signal SIGTERM +[2024-12-13 17:16:41,961] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797067 closing signal SIGTERM +[2024-12-13 17:16:41,961] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797068 closing signal SIGTERM +[2024-12-13 17:16:42,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797061 closing signal SIGTERM +[2024-12-13 17:16:42,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797062 closing signal SIGTERM +[2024-12-13 17:16:42,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797063 closing signal SIGTERM +[2024-12-13 17:16:42,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797064 closing signal SIGTERM +[2024-12-13 17:16:42,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797065 closing signal SIGTERM +[2024-12-13 17:16:42,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797066 closing signal SIGTERM +[2024-12-13 17:16:42,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797067 closing signal SIGTERM +[2024-12-13 17:16:42,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 797068 closing signal SIGTERM diff --git a/logs/13418046-bench_stress_test.out b/logs/13418046-bench_stress_test.out new file mode 100644 index 0000000000000000000000000000000000000000..4237e47d8977c58a12d3e51b780550c1f73b0315 --- /dev/null +++ b/logs/13418046-bench_stress_test.out @@ -0,0 +1,3768 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-169-207 ++ export NODELIST=ip-26-0-169-207 ++ NODELIST=ip-26-0-169-207 +++ scontrol show hostnames ip-26-0-169-207 +++ head -n1 ++ export MASTER_NODE=ip-26-0-169-207 ++ MASTER_NODE=ip-26-0-169-207 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-169-207' +Master node: ip-26-0-169-207 ++ echo 'All nodes: ip-26-0-169-207' +All nodes: ip-26-0-169-207 ++ echo 'World size: 8' +World size: 8 ++ srun torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13418046 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-207:12356 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k.yaml +[2024-12-13 17:47:55,063] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-13 17:47:55,063] torch.distributed.run: [WARNING] +[2024-12-13 17:47:55,063] torch.distributed.run: [WARNING] ***************************************** +[2024-12-13 17:47:55,063] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-13 17:47:55,063] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-169-207:810655:810655 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:810655:810655 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:810655:810655 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:810655:810655 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:810655:810655 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:810660:810660 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:810658:810658 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:810661:810661 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:810660:810660 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:810658:810658 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:810661:810661 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:810658:810658 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:810660:810660 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:810661:810661 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:810659:810659 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:810660:810660 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:810658:810658 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:810658:810658 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:810660:810660 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:810661:810661 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:810661:810661 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:810659:810659 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:810656:810656 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:810656:810656 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:810659:810659 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:810656:810656 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:810659:810659 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:810659:810659 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:810656:810656 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:810656:810656 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:810657:810657 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:810662:810662 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:810657:810657 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:810662:810662 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-207:810657:810657 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:810662:810662 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:810657:810657 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:810657:810657 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:810662:810662 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:810662:810662 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/162 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/149 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:810658:810778 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:810658:810778 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:810659:810782 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:810659:810782 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:810657:810783 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:810657:810783 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:810655:810781 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:810655:810781 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:810662:810784 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:810662:810784 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:810656:810779 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:810656:810779 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:810661:810780 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:810661:810780 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:810660:810777 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:810660:810777 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:810657:810783 [2] NCCL INFO comm 0x9eddce0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe5a41d427b2ea56b - Init START +ip-26-0-169-207:810656:810779 [1] NCCL INFO comm 0x97dc9e0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe5a41d427b2ea56b - Init START +ip-26-0-169-207:810662:810784 [7] NCCL INFO comm 0x9afaa30 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe5a41d427b2ea56b - Init START +ip-26-0-169-207:810661:810780 [6] NCCL INFO comm 0x849d150 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe5a41d427b2ea56b - Init START +ip-26-0-169-207:810660:810777 [5] NCCL INFO comm 0x895b1f0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe5a41d427b2ea56b - Init START +ip-26-0-169-207:810658:810778 [3] NCCL INFO comm 0x8c1e4a0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe5a41d427b2ea56b - Init START +ip-26-0-169-207:810655:810781 [0] NCCL INFO comm 0x9fa1710 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe5a41d427b2ea56b - Init START +ip-26-0-169-207:810659:810782 [4] NCCL INFO comm 0x9993e80 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe5a41d427b2ea56b - Init START +ip-26-0-169-207:810658:810778 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810657:810783 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810655:810781 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810661:810780 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810661:810780 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:810660:810777 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810660:810777 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:810659:810782 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810659:810782 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:810655:810781 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:810662:810784 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810662:810784 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:810657:810783 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:810658:810778 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:810656:810779 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:810657:810783 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:810658:810778 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:810656:810779 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:810658:810778 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:810658:810778 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810660:810777 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:810662:810784 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:810660:810777 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810662:810784 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810659:810782 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:810659:810782 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810661:810780 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:810657:810783 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:810657:810783 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810656:810779 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:810656:810779 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810661:810780 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810781 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:810655:810781 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Connected all rings +ip-26-0-169-207:810655:810781 [0] NCCL INFO Connected all rings +ip-26-0-169-207:810656:810779 [1] NCCL INFO Connected all rings +ip-26-0-169-207:810658:810778 [3] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810784 [7] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810778 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810661:810780 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810777 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810781 [0] NCCL INFO Connected all trees +ip-26-0-169-207:810655:810781 [0] NCCL INFO NVLS comm 0x9fa1710 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810784 [7] NCCL INFO Connected all trees +ip-26-0-169-207:810662:810784 [7] NCCL INFO NVLS comm 0x9afaa30 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810779 [1] NCCL INFO Connected all trees +ip-26-0-169-207:810656:810779 [1] NCCL INFO NVLS comm 0x97dc9e0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810782 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810657:810783 [2] NCCL INFO Connected all trees +ip-26-0-169-207:810657:810783 [2] NCCL INFO NVLS comm 0x9eddce0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810661:810780 [6] NCCL INFO Connected all trees +ip-26-0-169-207:810661:810780 [6] NCCL INFO NVLS comm 0x849d150 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810660:810777 [5] NCCL INFO Connected all trees +ip-26-0-169-207:810660:810777 [5] NCCL INFO NVLS comm 0x895b1f0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810659:810782 [4] NCCL INFO Connected all trees +ip-26-0-169-207:810659:810782 [4] NCCL INFO NVLS comm 0x9993e80 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810658:810778 [3] NCCL INFO Connected all trees +ip-26-0-169-207:810658:810778 [3] NCCL INFO NVLS comm 0x8c1e4a0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810660:810777 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810660:810777 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810660:810777 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810660:810777 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810662:810784 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810662:810784 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810662:810784 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810662:810784 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810656:810779 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810656:810779 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810656:810779 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810656:810779 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810661:810780 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810659:810782 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810661:810780 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810661:810780 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810661:810780 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810659:810782 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810659:810782 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810659:810782 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810657:810783 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810655:810781 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810657:810783 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810655:810781 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810657:810783 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810657:810783 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810655:810781 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810655:810781 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810658:810778 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810658:810778 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810658:810778 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810658:810778 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810662:810784 [7] NCCL INFO comm 0x9afaa30 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe5a41d427b2ea56b - Init COMPLETE +ip-26-0-169-207:810660:810777 [5] NCCL INFO comm 0x895b1f0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe5a41d427b2ea56b - Init COMPLETE +ip-26-0-169-207:810658:810778 [3] NCCL INFO comm 0x8c1e4a0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe5a41d427b2ea56b - Init COMPLETE +ip-26-0-169-207:810661:810780 [6] NCCL INFO comm 0x849d150 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe5a41d427b2ea56b - Init COMPLETE +ip-26-0-169-207:810656:810779 [1] NCCL INFO comm 0x97dc9e0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe5a41d427b2ea56b - Init COMPLETE +ip-26-0-169-207:810657:810783 [2] NCCL INFO comm 0x9eddce0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe5a41d427b2ea56b - Init COMPLETE +ip-26-0-169-207:810659:810782 [4] NCCL INFO comm 0x9993e80 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe5a41d427b2ea56b - Init COMPLETE +ip-26-0-169-207:810655:810781 [0] NCCL INFO comm 0x9fa1710 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe5a41d427b2ea56b - Init COMPLETE +ip-26-0-169-207:810655:810850 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:810661:810851 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:810659:810852 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:810657:810853 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:810660:810854 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:810658:810855 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:810656:810856 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:810662:810857 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:810660:810854 [5] NCCL INFO comm 0x8b851f0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa8c42c5a942699ef - Init START +ip-26-0-169-207:810656:810856 [1] NCCL INFO comm 0x9a06570 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa8c42c5a942699ef - Init START +ip-26-0-169-207:810658:810855 [3] NCCL INFO comm 0x8e48890 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa8c42c5a942699ef - Init START +ip-26-0-169-207:810657:810853 [2] NCCL INFO comm 0xa107bc0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa8c42c5a942699ef - Init START +ip-26-0-169-207:810659:810852 [4] NCCL INFO comm 0x9bbdb00 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa8c42c5a942699ef - Init START +ip-26-0-169-207:810662:810857 [7] NCCL INFO comm 0x9d244c0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa8c42c5a942699ef - Init START +ip-26-0-169-207:810661:810851 [6] NCCL INFO comm 0x86c6ac0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa8c42c5a942699ef - Init START +ip-26-0-169-207:810655:810850 [0] NCCL INFO comm 0xa1cbcc0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa8c42c5a942699ef - Init START +ip-26-0-169-207:810658:810855 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810657:810853 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810659:810852 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810660:810854 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810656:810856 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810662:810857 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810661:810851 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810655:810850 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:810661:810851 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810661:810851 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:810660:810854 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810660:810854 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:810655:810850 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:810658:810855 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:810658:810855 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:810657:810853 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:810657:810853 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:810656:810856 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:810656:810856 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:810662:810857 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810662:810857 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:810659:810852 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810659:810852 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:810659:810852 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:810660:810854 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:810659:810852 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810660:810854 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810662:810857 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:810662:810857 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810657:810853 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810658:810855 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810656:810856 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810661:810851 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810661:810851 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810657:810853 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810658:810855 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810656:810856 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:810655:810850 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Connected all rings +ip-26-0-169-207:810660:810854 [5] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Connected all rings +ip-26-0-169-207:810657:810853 [2] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810854 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810856 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810852 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810855 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810661:810851 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810655:810850 [0] NCCL INFO Connected all trees +ip-26-0-169-207:810655:810850 [0] NCCL INFO NVLS comm 0xa1cbcc0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810853 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810857 [7] NCCL INFO Connected all trees +ip-26-0-169-207:810662:810857 [7] NCCL INFO NVLS comm 0x9d244c0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810659:810852 [4] NCCL INFO Connected all trees +ip-26-0-169-207:810659:810852 [4] NCCL INFO NVLS comm 0x9bbdb00 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810661:810851 [6] NCCL INFO Connected all trees +ip-26-0-169-207:810661:810851 [6] NCCL INFO NVLS comm 0x86c6ac0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810660:810854 [5] NCCL INFO Connected all trees +ip-26-0-169-207:810660:810854 [5] NCCL INFO NVLS comm 0x8b851f0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810656:810856 [1] NCCL INFO Connected all trees +ip-26-0-169-207:810656:810856 [1] NCCL INFO NVLS comm 0x9a06570 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810658:810855 [3] NCCL INFO Connected all trees +ip-26-0-169-207:810658:810855 [3] NCCL INFO NVLS comm 0x8e48890 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810657:810853 [2] NCCL INFO Connected all trees +ip-26-0-169-207:810657:810853 [2] NCCL INFO NVLS comm 0xa107bc0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810655:810850 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810655:810850 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810655:810850 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810655:810850 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810659:810852 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810659:810852 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810659:810852 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810659:810852 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810662:810857 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810662:810857 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810662:810857 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810662:810857 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810660:810854 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810660:810854 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810660:810854 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810660:810854 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810656:810856 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810656:810856 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810656:810856 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810656:810856 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810657:810853 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810657:810853 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810657:810853 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810657:810853 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810658:810855 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810658:810855 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810658:810855 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810658:810855 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810661:810851 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810661:810851 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810661:810851 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810661:810851 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810660:810854 [5] NCCL INFO comm 0x8b851f0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa8c42c5a942699ef - Init COMPLETE +ip-26-0-169-207:810655:810850 [0] NCCL INFO comm 0xa1cbcc0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa8c42c5a942699ef - Init COMPLETE +ip-26-0-169-207:810662:810857 [7] NCCL INFO comm 0x9d244c0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa8c42c5a942699ef - Init COMPLETE +ip-26-0-169-207:810659:810852 [4] NCCL INFO comm 0x9bbdb00 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa8c42c5a942699ef - Init COMPLETE +ip-26-0-169-207:810661:810851 [6] NCCL INFO comm 0x86c6ac0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa8c42c5a942699ef - Init COMPLETE +ip-26-0-169-207:810658:810855 [3] NCCL INFO comm 0x8e48890 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa8c42c5a942699ef - Init COMPLETE +ip-26-0-169-207:810656:810856 [1] NCCL INFO comm 0x9a06570 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa8c42c5a942699ef - Init COMPLETE +ip-26-0-169-207:810657:810853 [2] NCCL INFO comm 0xa107bc0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa8c42c5a942699ef - Init COMPLETE +ip-26-0-169-207:810655:810883 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:810656:810884 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:810662:810888 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:810660:810887 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:810659:810889 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:810661:810890 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:810657:810886 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:810658:810885 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:810658:810885 [3] NCCL INFO comm 0x8e5ce70 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa1537531c24d3ab2 - Init START +ip-26-0-169-207:810657:810886 [2] NCCL INFO comm 0xa11c300 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa1537531c24d3ab2 - Init START +ip-26-0-169-207:810662:810888 [7] NCCL INFO comm 0x9d389f0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa1537531c24d3ab2 - Init START +ip-26-0-169-207:810659:810889 [4] NCCL INFO comm 0x9bd2250 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa1537531c24d3ab2 - Init START +ip-26-0-169-207:810660:810887 [5] NCCL INFO comm 0x8b99810 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa1537531c24d3ab2 - Init START +ip-26-0-169-207:810656:810884 [1] NCCL INFO comm 0x9a1acb0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa1537531c24d3ab2 - Init START +ip-26-0-169-207:810655:810883 [0] NCCL INFO comm 0xa1e01c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa1537531c24d3ab2 - Init START +ip-26-0-169-207:810661:810890 [6] NCCL INFO comm 0x86db200 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa1537531c24d3ab2 - Init START +ip-26-0-169-207:810658:810885 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810659:810889 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810657:810886 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810660:810887 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810656:810884 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810661:810890 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810662:810888 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810655:810883 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:810658:810885 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:810658:810885 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:810660:810887 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810660:810887 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:810659:810889 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810659:810889 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:810661:810890 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810657:810886 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:810657:810886 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:810661:810890 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:810662:810888 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810662:810888 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:810656:810884 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:810656:810884 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:810655:810883 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810662:810888 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810656:810884 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:810662:810888 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810656:810884 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:810660:810887 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-207:810660:810887 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810659:810889 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:810655:810883 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810659:810889 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810657:810886 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:810661:810890 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-207:810661:810890 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810658:810885 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-207:810658:810885 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810657:810886 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Connected all rings +ip-26-0-169-207:810658:810885 [3] NCCL INFO Connected all rings +ip-26-0-169-207:810657:810886 [2] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810655:810883 [0] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810660:810887 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810661:810890 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810657:810886 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810658:810885 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810656:810884 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810659:810889 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-207:810662:810888 [7] NCCL INFO Connected all trees +ip-26-0-169-207:810662:810888 [7] NCCL INFO NVLS comm 0x9d389f0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810655:810883 [0] NCCL INFO Connected all trees +ip-26-0-169-207:810655:810883 [0] NCCL INFO NVLS comm 0xa1e01c0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810661:810890 [6] NCCL INFO Connected all trees +ip-26-0-169-207:810661:810890 [6] NCCL INFO NVLS comm 0x86db200 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810660:810887 [5] NCCL INFO Connected all trees +ip-26-0-169-207:810660:810887 [5] NCCL INFO NVLS comm 0x8b99810 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810657:810886 [2] NCCL INFO Connected all trees +ip-26-0-169-207:810657:810886 [2] NCCL INFO NVLS comm 0xa11c300 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810656:810884 [1] NCCL INFO Connected all trees +ip-26-0-169-207:810656:810884 [1] NCCL INFO NVLS comm 0x9a1acb0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810658:810885 [3] NCCL INFO Connected all trees +ip-26-0-169-207:810658:810885 [3] NCCL INFO NVLS comm 0x8e5ce70 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810659:810889 [4] NCCL INFO Connected all trees +ip-26-0-169-207:810659:810889 [4] NCCL INFO NVLS comm 0x9bd2250 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-207:810660:810887 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810660:810887 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810660:810887 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810660:810887 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810656:810884 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810656:810884 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810656:810884 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810656:810884 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810658:810885 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810658:810885 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810658:810885 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810658:810885 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810662:810888 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810662:810888 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810662:810888 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810662:810888 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810659:810889 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810659:810889 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810659:810889 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810659:810889 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810661:810890 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810661:810890 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810661:810890 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810661:810890 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810657:810886 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810657:810886 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810657:810886 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810657:810886 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810655:810883 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810655:810883 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810655:810883 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:810655:810883 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810657:810886 [2] NCCL INFO comm 0xa11c300 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa1537531c24d3ab2 - Init COMPLETE +ip-26-0-169-207:810659:810889 [4] NCCL INFO comm 0x9bd2250 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa1537531c24d3ab2 - Init COMPLETE +ip-26-0-169-207:810655:810883 [0] NCCL INFO comm 0xa1e01c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa1537531c24d3ab2 - Init COMPLETE +ip-26-0-169-207:810656:810884 [1] NCCL INFO comm 0x9a1acb0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa1537531c24d3ab2 - Init COMPLETE +ip-26-0-169-207:810662:810888 [7] NCCL INFO comm 0x9d389f0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa1537531c24d3ab2 - Init COMPLETE +ip-26-0-169-207:810658:810885 [3] NCCL INFO comm 0x8e5ce70 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa1537531c24d3ab2 - Init COMPLETE +ip-26-0-169-207:810661:810890 [6] NCCL INFO comm 0x86db200 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa1537531c24d3ab2 - Init COMPLETE +ip-26-0-169-207:810660:810887 [5] NCCL INFO comm 0x8b99810 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa1537531c24d3ab2 - Init COMPLETE +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config: +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Config(general=GeneralArgs(project='debug', +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: run='1.14G_dp4_tp2_pp1_acc256_mbs2_seq2048_zero1_tpmodeRED_vocab32k', +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: step=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: consumed_train_samples=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: benchmark_csv_path=PosixPath('benchmark/results/bench_final.csv'), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ignore_sanity_checks=True), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: parallelism=ParallelismArgs(dp=4, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp=1, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp=2, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pp_engine=, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_mode=, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_linear_async_communication=True, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: recompute_layer=False, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tp_recompute_allgather=True, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: expert_parallel_size=1), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: init_method=RandomInit(std=0.02), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: dtype=torch.bfloat16, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: make_vocab_size_divisible_by=1, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: ddp_bucket_cap_mb=25), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_revision=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokenizer_max_length=None), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoint_interval=10000, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_initial_state=False, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: save_final_state=False, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: resume_checkpoint_path=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: checkpoints_path_is_shared_file_system=False), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: logging=LoggingArgs(log_level='info', +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: log_level_replica='info', +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration_step_info_interval=1), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tokens=TokensArgs(sequence_length=2048, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: train_steps=100, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: micro_batch_size=2, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: batch_accumulation_per_replica=256, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: val_check_interval=100, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_val_batches=0, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: limit_test_batches=0), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta1=0.9, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: adam_beta2=0.95, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: torch_adam_is_fused=True, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: name='adamW'), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: zero_stage=1, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: weight_decay=0.01, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: clip_grad=1.0, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: accumulate_grad_in_fp32=True, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_steps=2, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_warmup_style='linear', +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_style='cosine', +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_steps=13, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lr_decay_starting_step=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: min_decay_lr=1e-05)), +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: start_training_step=1, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: data=DataArgs(dataset=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: seed=42, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_loading_workers=1))], +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: profiler=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: lighteval=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: s3_upload=None) +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Model Config: +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: LlamaConfig(bos_token_id=0, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: eos_token_id=0, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_act='silu', +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: hidden_size=2048, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: initializer_range=0.02, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: intermediate_size=8192, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: is_llama_config=True, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: max_position_embeddings=2048, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_attention_heads=32, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_hidden_layers=16, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: num_key_value_heads=8, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pad_token_id=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: pretraining_tp=1, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rms_norm_eps=1e-05, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_scaling=None, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_theta=10000.0, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: rope_interleaved=False, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: tie_word_embeddings=True, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: use_cache=True, +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: vocab_size=32768) +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Building model.. +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Initialize RoPE Theta = 10000.0 +12/13/2024 17:48:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:810659:810938 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:810658:810943 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:810656:810942 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:810662:810944 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:810660:810945 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:810655:810940 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:810657:810941 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:810661:810939 [6] NCCL INFO comm 0xa4e03b0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9e4d98aada75fad2 - Init START +ip-26-0-169-207:810662:810944 [7] NCCL INFO comm 0xbb3a5f0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9e4d98aada75fad2 - Init START +ip-26-0-169-207:810659:810938 [4] NCCL INFO comm 0xb9d6ab0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8df4ecd33856b7a - Init START +ip-26-0-169-207:810660:810945 [5] NCCL INFO comm 0xa999ea0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8df4ecd33856b7a - Init START +ip-26-0-169-207:810655:810940 [0] NCCL INFO comm 0xbfe21f0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9bffdd2d014ff331 - Init START +ip-26-0-169-207:810657:810941 [2] NCCL INFO comm 0xbf1d5b0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc97232f83e7e5590 - Init START +ip-26-0-169-207:810656:810942 [1] NCCL INFO comm 0xb81d030 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9bffdd2d014ff331 - Init START +ip-26-0-169-207:810658:810943 [3] NCCL INFO comm 0xac5d9f0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc97232f83e7e5590 - Init START +ip-26-0-169-207:810661:810939 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810659:810938 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810655:810940 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:810657:810941 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810662:810944 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810660:810945 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810656:810942 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810658:810943 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810656:810942 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:810655:810940 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:810656:810942 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:810656:810942 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:810655:810940 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:810662:810944 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:810662:810944 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:810662:810944 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:810661:810939 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810660:810945 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:810660:810945 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:810660:810945 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:810659:810938 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810658:810943 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-207:810658:810943 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-207:810658:810943 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-207:810657:810941 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810655:810940 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810662:810944 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810658:810943 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810656:810942 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810660:810945 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810661:810939 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810659:810938 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-207:810657:810941 [2] NCCL INFO Connected all rings +ip-26-0-169-207:810657:810941 [2] NCCL INFO Connected all trees +ip-26-0-169-207:810657:810941 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810657:810941 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:810657:810941 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810655:810940 [0] NCCL INFO Connected all rings +ip-26-0-169-207:810655:810940 [0] NCCL INFO Connected all trees +ip-26-0-169-207:810655:810940 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810655:810940 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:810655:810940 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810658:810943 [3] NCCL INFO Connected all rings +ip-26-0-169-207:810658:810943 [3] NCCL INFO Connected all trees +ip-26-0-169-207:810658:810943 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810658:810943 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:810658:810943 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810656:810942 [1] NCCL INFO Connected all rings +ip-26-0-169-207:810656:810942 [1] NCCL INFO Connected all trees +ip-26-0-169-207:810656:810942 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810656:810942 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:810656:810942 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810658:810943 [3] NCCL INFO comm 0xac5d9f0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc97232f83e7e5590 - Init COMPLETE +ip-26-0-169-207:810657:810941 [2] NCCL INFO comm 0xbf1d5b0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc97232f83e7e5590 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:810658:810962 [3] NCCL INFO comm 0xac71830 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x69add772297b5df7 - Init START +ip-26-0-169-207:810658:810962 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810661:810939 [6] NCCL INFO Connected all rings +ip-26-0-169-207:810661:810939 [6] NCCL INFO Connected all trees +ip-26-0-169-207:810661:810939 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810661:810939 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:810661:810939 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810660:810945 [5] NCCL INFO Connected all rings +ip-26-0-169-207:810660:810945 [5] NCCL INFO Connected all trees +ip-26-0-169-207:810660:810945 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810660:810945 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:810660:810945 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810655:810940 [0] NCCL INFO comm 0xbfe21f0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9bffdd2d014ff331 - Init COMPLETE +ip-26-0-169-207:810656:810942 [1] NCCL INFO comm 0xb81d030 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9bffdd2d014ff331 - Init COMPLETE +ip-26-0-169-207:810662:810944 [7] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810944 [7] NCCL INFO Connected all trees +ip-26-0-169-207:810662:810944 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810662:810944 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:810662:810944 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810655:810966 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:810655:810966 [0] NCCL INFO comm 0xbff6720 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x69e0afdeeee76dda - Init START +ip-26-0-169-207:810655:810966 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:810657:810969 [2] NCCL INFO comm 0xbf311e0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa84699ae7bcd5576 - Init START +ip-26-0-169-207:810657:810969 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810659:810938 [4] NCCL INFO Connected all rings +ip-26-0-169-207:810659:810938 [4] NCCL INFO Connected all trees +ip-26-0-169-207:810659:810938 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810659:810938 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:810659:810938 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:810661:810939 [6] NCCL INFO comm 0xa4e03b0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9e4d98aada75fad2 - Init COMPLETE +ip-26-0-169-207:810662:810944 [7] NCCL INFO comm 0xbb3a5f0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9e4d98aada75fad2 - Init COMPLETE +ip-26-0-169-207:810656:810972 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:810656:810972 [1] NCCL INFO comm 0xb830c60 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7173eacedc4b2fba - Init START +ip-26-0-169-207:810656:810972 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:810662:810975 [7] NCCL INFO comm 0xbb4dbd0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa3febba4c83d7428 - Init START +ip-26-0-169-207:810662:810975 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:810661:810976 [6] NCCL INFO comm 0xa4f4100 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x135251ec088f8f07 - Init START +ip-26-0-169-207:810661:810976 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810660:810945 [5] NCCL INFO comm 0xa999ea0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8df4ecd33856b7a - Init COMPLETE +ip-26-0-169-207:810659:810938 [4] NCCL INFO comm 0xb9d6ab0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8df4ecd33856b7a - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:810660:810980 [5] NCCL INFO comm 0xa9adbc0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6505f38a47439004 - Init START +ip-26-0-169-207:810659:810979 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:810660:810980 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810659:810979 [4] NCCL INFO comm 0xb9ea8b0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x841dbcc5d11bf7c9 - Init START +ip-26-0-169-207:810659:810979 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:810658:810962 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:810658:810962 [3] NCCL INFO Connected all rings +ip-26-0-169-207:810658:810962 [3] NCCL INFO Connected all trees +ip-26-0-169-207:810658:810962 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810655:810966 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:810655:810966 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:810655:810966 [0] NCCL INFO Connected all rings +ip-26-0-169-207:810655:810966 [0] NCCL INFO Connected all trees +ip-26-0-169-207:810655:810966 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810657:810969 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:810657:810969 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:810657:810969 [2] NCCL INFO Connected all rings +ip-26-0-169-207:810657:810969 [2] NCCL INFO Connected all trees +ip-26-0-169-207:810657:810969 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810656:810972 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:810656:810972 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:810656:810972 [1] NCCL INFO Connected all rings +ip-26-0-169-207:810656:810972 [1] NCCL INFO Connected all trees +ip-26-0-169-207:810656:810972 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810658:810962 [3] NCCL INFO comm 0xac71830 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x69add772297b5df7 - Init COMPLETE +ip-26-0-169-207:810659:810979 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:810659:810979 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:810659:810979 [4] NCCL INFO Connected all rings +ip-26-0-169-207:810659:810979 [4] NCCL INFO Connected all trees +ip-26-0-169-207:810659:810979 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810655:810966 [0] NCCL INFO comm 0xbff6720 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x69e0afdeeee76dda - Init COMPLETE +ip-26-0-169-207:810661:810976 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:810661:810976 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:810661:810976 [6] NCCL INFO Connected all rings +ip-26-0-169-207:810661:810976 [6] NCCL INFO Connected all trees +ip-26-0-169-207:810661:810976 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810662:810975 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:810662:810975 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:810662:810975 [7] NCCL INFO Connected all rings +ip-26-0-169-207:810662:810975 [7] NCCL INFO Connected all trees +ip-26-0-169-207:810662:810975 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810657:810969 [2] NCCL INFO comm 0xbf311e0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa84699ae7bcd5576 - Init COMPLETE +ip-26-0-169-207:810656:810972 [1] NCCL INFO comm 0xb830c60 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7173eacedc4b2fba - Init COMPLETE +12/13/2024 17:48:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Total number of parameters: 1.04G (1984.26MiB) +12/13/2024 17:48:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +12/13/2024 17:48:33 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: Local number of parameters: 520M (992.13MiB) +ip-26-0-169-207:810660:810980 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-207:810660:810980 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-207:810660:810980 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:810659:810979 [4] NCCL INFO comm 0xb9ea8b0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x841dbcc5d11bf7c9 - Init COMPLETE +12/13/2024 17:48:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +12/13/2024 17:48:34 [INFO|DP=0|PP=0|TP=1|ip-26-0-169-207]: [After model building] Memory usage: 992.15MiB. Peak allocated: 5632.00MiB Peak reserved: 16898.00MiB +ip-26-0-169-207:810660:810980 [5] NCCL INFO Connected all rings +ip-26-0-169-207:810660:810980 [5] NCCL INFO Connected all trees +ip-26-0-169-207:810660:810980 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +12/13/2024 17:48:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: No checkpoint path provided. +12/13/2024 17:48:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Parametrizing model parameters using StandardParametrizator +ip-26-0-169-207:810655:810999 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:810657:811000 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:810656:811003 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:810658:811002 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:810661:810976 [6] NCCL INFO comm 0xa4f4100 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x135251ec088f8f07 - Init COMPLETE +ip-26-0-169-207:810662:810975 [7] NCCL INFO comm 0xbb4dbd0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa3febba4c83d7428 - Init COMPLETE +ip-26-0-169-207:810660:810980 [5] NCCL INFO comm 0xa9adbc0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6505f38a47439004 - Init COMPLETE +ip-26-0-169-207:810661:811004 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:810662:811005 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:810660:811007 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:810659:811006 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:810660:811007 [5] NCCL INFO comm 0xa9b3140 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd040828515487813 - Init START +ip-26-0-169-207:810656:811003 [1] NCCL INFO comm 0xb835b80 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd040828515487813 - Init START +ip-26-0-169-207:810658:811002 [3] NCCL INFO comm 0xac76450 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd040828515487813 - Init START +ip-26-0-169-207:810662:811005 [7] NCCL INFO comm 0xbb53160 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd040828515487813 - Init START +ip-26-0-169-207:810662:811005 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810660:811007 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810658:811002 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810656:811003 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810661:811004 [6] NCCL INFO comm 0xa4f9200 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x424354d35aca987b - Init START +ip-26-0-169-207:810655:810999 [0] NCCL INFO comm 0xbffb8e0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x424354d35aca987b - Init START +ip-26-0-169-207:810659:811006 [4] NCCL INFO comm 0xb9ef9d0 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x424354d35aca987b - Init START +ip-26-0-169-207:810657:811000 [2] NCCL INFO comm 0xbf36270 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x424354d35aca987b - Init START +ip-26-0-169-207:810657:811000 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810661:811004 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810659:811006 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/149 +ip-26-0-169-207:810655:810999 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/162 +ip-26-0-169-207:810660:811007 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810660:811007 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-207:810662:811005 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810662:811005 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-207:810659:811006 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810659:811006 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:810655:810999 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-207:810661:811004 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:810661:811004 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:810656:811003 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-207:810658:811002 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:810658:811002 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-207:810657:811000 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:810657:811000 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:810658:811002 [3] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:810658:811002 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:810660:811007 [5] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:810662:811005 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:810660:811007 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810662:811005 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:810656:811003 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-207:810657:811000 [2] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-207:810657:811000 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-207:810661:811004 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-207:810661:811004 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-207:810659:811006 [4] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-207:810659:811006 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-207:810655:810999 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 04/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 05/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 06/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 04/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 07/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 05/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 08/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 06/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 09/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 07/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 10/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 08/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 11/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 09/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 04/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 04/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 12/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 05/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 13/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 10/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 05/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 06/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 14/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 04/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 06/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 15/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 07/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 05/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 07/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 08/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 16/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 06/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 08/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 17/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 09/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 07/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 09/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 11/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 18/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 10/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 08/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 10/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 12/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 19/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 11/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 09/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 11/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 13/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 20/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 12/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 10/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 12/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 14/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 21/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 13/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 22/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 11/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 15/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 14/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 13/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 16/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 17/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 12/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 18/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 15/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 13/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 16/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 19/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 14/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 20/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 17/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 23/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 14/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 15/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 18/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 21/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 15/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 16/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 19/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 22/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 16/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 17/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 20/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 23/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 17/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 18/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 21/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 18/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 22/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 19/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 19/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 23/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 20/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 20/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 21/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 21/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 22/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 22/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 23/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 23/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Connected all rings +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810656:811003 [1] NCCL INFO Connected all rings +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Connected all rings +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Connected all rings +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810655:810999 [0] NCCL INFO Connected all rings +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Connected all rings +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Connected all rings +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Connected all rings +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 02/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 03/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 04/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 05/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 06/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 07/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 08/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 09/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 10/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 04/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 11/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 05/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 12/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 04/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 06/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 13/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 07/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 05/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 14/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 08/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 06/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 15/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 09/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 07/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 16/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 10/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 08/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 17/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 11/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 09/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 18/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 12/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 10/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 19/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 11/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 13/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 20/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 12/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 14/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 21/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 15/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 13/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 22/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 14/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 16/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810661:811004 [6] NCCL INFO Channel 23/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 15/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 17/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 16/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 18/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 17/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 19/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 18/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 19/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 20/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 20/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 21/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 21/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 22/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810657:811000 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-169-207:810660:811007 [5] NCCL INFO Channel 23/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 22/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810659:811006 [4] NCCL INFO Channel 23/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810658:811002 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-169-207:810662:811005 [7] NCCL INFO Connected all trees +ip-26-0-169-207:810662:811005 [7] NCCL INFO NVLS comm 0xbb53160 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:810661:811004 [6] NCCL INFO Connected all trees +ip-26-0-169-207:810661:811004 [6] NCCL INFO NVLS comm 0xa4f9200 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:810655:810999 [0] NCCL INFO Connected all trees +ip-26-0-169-207:810655:810999 [0] NCCL INFO NVLS comm 0xbffb8e0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:810656:811003 [1] NCCL INFO Connected all trees +ip-26-0-169-207:810656:811003 [1] NCCL INFO NVLS comm 0xb835b80 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:810657:811000 [2] NCCL INFO Connected all trees +ip-26-0-169-207:810657:811000 [2] NCCL INFO NVLS comm 0xbf36270 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:810660:811007 [5] NCCL INFO Connected all trees +ip-26-0-169-207:810660:811007 [5] NCCL INFO NVLS comm 0xa9b3140 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:810658:811002 [3] NCCL INFO Connected all trees +ip-26-0-169-207:810658:811002 [3] NCCL INFO NVLS comm 0xac76450 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:810659:811006 [4] NCCL INFO Connected all trees +ip-26-0-169-207:810659:811006 [4] NCCL INFO NVLS comm 0xb9ef9d0 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-207:810660:811007 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810660:811007 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810656:811003 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810660:811007 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:810660:811007 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810656:811003 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810656:811003 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:810656:811003 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810662:811005 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810662:811005 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810662:811005 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:810662:811005 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810658:811002 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810658:811002 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810658:811002 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:810658:811002 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810656:811003 [1] NCCL INFO comm 0xb835b80 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd040828515487813 - Init COMPLETE +ip-26-0-169-207:810660:811007 [5] NCCL INFO comm 0xa9b3140 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd040828515487813 - Init COMPLETE +ip-26-0-169-207:810658:811002 [3] NCCL INFO comm 0xac76450 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd040828515487813 - Init COMPLETE +ip-26-0-169-207:810662:811005 [7] NCCL INFO comm 0xbb53160 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd040828515487813 - Init COMPLETE +ip-26-0-169-207:810659:811006 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810659:811006 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810659:811006 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:810659:811006 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810661:811004 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810661:811004 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810661:811004 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:810661:811004 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810655:810999 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810655:810999 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810655:810999 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:810655:810999 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810657:811000 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-207:810657:811000 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-207:810657:811000 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:810657:811000 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-207:810659:811006 [4] NCCL INFO comm 0xb9ef9d0 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x424354d35aca987b - Init COMPLETE +ip-26-0-169-207:810655:810999 [0] NCCL INFO comm 0xbffb8e0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x424354d35aca987b - Init COMPLETE +ip-26-0-169-207:810657:811000 [2] NCCL INFO comm 0xbf36270 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x424354d35aca987b - Init COMPLETE +ip-26-0-169-207:810661:811004 [6] NCCL INFO comm 0xa4f9200 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x424354d35aca987b - Init COMPLETE +12/13/2024 17:48:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Optimizer Building] Using LearningRateForSP as learning rate +12/13/2024 17:48:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] Size of optimizer params per rank: +12/13/2024 17:48:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 0 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 17:48:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 1 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 17:48:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 2 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 17:48:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [ZeRO sharding] DP Rank 3 has 130M out of 520M (25.00%) params' optimizer states +12/13/2024 17:48:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/13/2024 17:48:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Using dummy data generator +12/13/2024 17:48:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Training Plan] There are 1 training stages +12/13/2024 17:48:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Stage Stable Training Stage] start from step 1 +12/13/2024 17:48:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: +12/13/2024 17:48:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: [Start training] datetime: 2024-12-13 17:48:38.654817 | mbs: 2 | grad_accum: 256 | global_batch_size: 2048 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory. +wandb: WARNING Path /fsx/nouamane/.cache/wandb/wandb/ wasn't writable, using system temp directory +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/13/2024 17:48:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Waiting for 14 seconds +12/13/2024 17:49:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/13/2024 17:49:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3472.48MiB. Peak allocated 5632.00MiB. Peak reserved: 18884.00MiB +12/13/2024 17:49:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 3572.67MiB. Peak allocated 6985.16MiB. Peak reserved: 7434.00MiB +12/13/2024 17:49:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 20.1K | tokens_per_sec: 209K | tokens_per_sec_per_gpu: 26.1K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.00015 | model_tflops_per_gpu: 184 | hardware_tflops_per_gpu: 184 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 9.62G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 17:49:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.85MiB. Peak allocated 6820.14MiB. Peak reserved: 9178.00MiB +12/13/2024 17:49:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.97MiB. Peak allocated 8000.47MiB. Peak reserved: 9434.00MiB +12/13/2024 17:49:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 18.2K | tokens_per_sec: 230K | tokens_per_sec_per_gpu: 28.8K | global_batch_size: 2.05K | lm_loss: 10.8 | lr: 0.0003 | model_tflops_per_gpu: 203 | hardware_tflops_per_gpu: 203 | grad_norm: 0.431 | cuda_memory_allocated: 4.81G | cuda_max_memory_reserved: 10.9G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.7G | hd_free_memory_tb: 242G +12/13/2024 17:49:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-207]: Memory usage: 4587.85MiB. Peak allocated 6820.26MiB. Peak reserved: 10428.00MiB +slurmstepd: error: *** JOB 13418046 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T17:49:52 DUE TO TIME LIMIT *** +slurmstepd: error: *** STEP 13418046.0 ON ip-26-0-169-207 CANCELLED AT 2024-12-13T17:49:52 DUE TO TIME LIMIT *** +[2024-12-13 17:49:52,532] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-13 17:49:52,532] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810655 closing signal SIGTERM +[2024-12-13 17:49:52,532] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810656 closing signal SIGTERM +[2024-12-13 17:49:52,532] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810657 closing signal SIGTERM +[2024-12-13 17:49:52,532] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810658 closing signal SIGTERM +[2024-12-13 17:49:52,533] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810659 closing signal SIGTERM +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-13 17:49:52,534] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810660 closing signal SIGTERM +[2024-12-13 17:49:52,535] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810661 closing signal SIGTERM +[2024-12-13 17:49:52,535] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810662 closing signal SIGTERM +[2024-12-13 17:49:52,688] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810655 closing signal SIGTERM +[2024-12-13 17:49:52,688] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810656 closing signal SIGTERM +[2024-12-13 17:49:52,689] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810657 closing signal SIGTERM +[2024-12-13 17:49:52,689] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810658 closing signal SIGTERM +[2024-12-13 17:49:52,689] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810659 closing signal SIGTERM +[2024-12-13 17:49:52,689] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810660 closing signal SIGTERM +[2024-12-13 17:49:52,689] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810661 closing signal SIGTERM +[2024-12-13 17:49:52,689] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 810662 closing signal SIGTERM diff --git a/logs/13440897-bench_80G_dp16_tp8_pp1_acc16_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/13440897-bench_80G_dp16_tp8_pp1_acc16_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..f51fb20797bf9177f8d6dcf2635909eca2bd9307 --- /dev/null +++ b/logs/13440897-bench_80G_dp16_tp8_pp1_acc16_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,3678 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-160-[100,103,225,242],ip-26-0-161-[78,103,123],ip-26-0-165-[59,131,164,202,213],ip-26-0-172-[116,142,147,252]' ++ export 'NODELIST=ip-26-0-160-100 +ip-26-0-160-103 +ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-161-103 +ip-26-0-161-123 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252' ++ NODELIST='ip-26-0-160-100 +ip-26-0-160-103 +ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-161-103 +ip-26-0-161-123 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-160-[100,103,225,242],ip-26-0-161-[78,103,123],ip-26-0-165-[59,131,164,202,213],ip-26-0-172-[116,142,147,252]' ++ export MASTER_NODE=ip-26-0-160-100 ++ MASTER_NODE=ip-26-0-160-100 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=16 ++ NNODES=16 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=128 ++ WORLD_SIZE=128 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-160-100' +Master node: ip-26-0-160-100 ++ echo 'All nodes: ip-26-0-160-100 +ip-26-0-160-103 +ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-161-103 +ip-26-0-161-123 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252' +All nodes: ip-26-0-160-100 +ip-26-0-160-103 +ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-161-103 +ip-26-0-161-123 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 ++ echo 'World size: 128' +World size: 128 ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=16 --nproc_per_node=8 --rdzv_id=13440897 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-100:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_80G_dp16_tp8_pp1_acc16_mbs1_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-14 03:58:22,906] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,906] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,908] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,915] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,916] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,932] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,931] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,944] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,906] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,906] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,906] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,906] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,907] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,907] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,907] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,907] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,908] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,908] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,908] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,908] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,915] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,915] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,915] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,915] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,916] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,916] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,916] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,916] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,968] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,968] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,931] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,931] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,931] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,931] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,932] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,932] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,932] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,932] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,978] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,978] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,944] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,944] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,944] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,944] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,995] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,969] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,969] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,969] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,969] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:23,010] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,968] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,968] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,968] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,968] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,979] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,979] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,979] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,979] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,978] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,978] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,978] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,978] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:23,044] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:22,995] torch.distributed.run: [WARNING] +[2024-12-14 03:58:22,995] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:22,995] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:22,995] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:23,010] torch.distributed.run: [WARNING] +[2024-12-14 03:58:23,010] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:23,010] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:23,010] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:23,075] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:58:23,044] torch.distributed.run: [WARNING] +[2024-12-14 03:58:23,044] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:23,044] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:23,044] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:23,075] torch.distributed.run: [WARNING] +[2024-12-14 03:58:23,075] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:58:23,075] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:58:23,075] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +12/14/2024 03:58:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Measuring inter-GPU and intra-node bandwidth... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Bandwidth measurement complete. Time taken: 18.25 seconds +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Config: +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Config(general=GeneralArgs(project='debug', +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: run='80G_dp16_tp8_pp1_acc16_mbs1_seq4096_zero0_tpmodeRED_vocab131k', +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: seed=42, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: step=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: consumed_train_samples=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: ignore_sanity_checks=True), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: parallelism=ParallelismArgs(dp=16, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pp=1, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp=8, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pp_engine=, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp_mode=, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp_linear_async_communication=True, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: recompute_layer=False, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp_recompute_allgather=True, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: expert_parallel_size=1), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: eos_token_id=0, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_act='silu', +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_size=8192, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: initializer_range=0.02, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: intermediate_size=28672, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: is_llama_config=True, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: max_position_embeddings=4096, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_attention_heads=64, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_hidden_layers=80, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_key_value_heads=64, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pad_token_id=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pretraining_tp=1, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rms_norm_eps=1e-05, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_scaling=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_theta=10000.0, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_interleaved=False, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tie_word_embeddings=False, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: use_cache=True, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: vocab_size=131072), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: init_method=RandomInit(std=0.02), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: dtype=torch.bfloat16, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: make_vocab_size_divisible_by=1, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: ddp_bucket_cap_mb=25), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokenizer_revision=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokenizer_max_length=None), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: checkpoint_interval=10000, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: save_initial_state=False, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: save_final_state=False, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: resume_checkpoint_path=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: checkpoints_path_is_shared_file_system=False), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: logging=LoggingArgs(log_level='info', +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: log_level_replica='info', +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: iteration_step_info_interval=1), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokens=TokensArgs(sequence_length=4096, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: train_steps=100, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: micro_batch_size=1, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: batch_accumulation_per_replica=16, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: val_check_interval=100, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: limit_val_batches=0, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: limit_test_batches=0), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: adam_beta1=0.9, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: adam_beta2=0.95, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: torch_adam_is_fused=True, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: name='adamW'), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: zero_stage=0, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: weight_decay=0.01, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: clip_grad=1.0, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: accumulate_grad_in_fp32=True, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_warmup_steps=2, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_warmup_style='linear', +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_decay_style='cosine', +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_decay_steps=13, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_decay_starting_step=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: min_decay_lr=1e-05)), +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: start_training_step=1, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: data=DataArgs(dataset=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: seed=42, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_loading_workers=1))], +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: profiler=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lighteval=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: s3_upload=None) +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Model Config: +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: LlamaConfig(bos_token_id=0, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: eos_token_id=0, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_act='silu', +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_size=8192, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: initializer_range=0.02, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: intermediate_size=28672, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: is_llama_config=True, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: max_position_embeddings=4096, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_attention_heads=64, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_hidden_layers=80, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_key_value_heads=64, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pad_token_id=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pretraining_tp=1, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rms_norm_eps=1e-05, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_scaling=None, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_theta=10000.0, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_interleaved=False, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tie_word_embeddings=False, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: use_cache=True, +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: vocab_size=131072) +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Building model.. +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Initialize RoPE Theta = 10000.0 +12/14/2024 03:59:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=4|ip-26-0-160-100]: Local number of parameters: 10G (19074.52MiB) +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=7|ip-26-0-160-100]: Local number of parameters: 10G (19074.52MiB) +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=5|ip-26-0-160-100]: Local number of parameters: 10G (19074.52MiB) +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-100]: Local number of parameters: 10G (19074.52MiB) +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-100]: Local number of parameters: 10G (19074.52MiB) +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Total number of parameters: 80G (152596.12MiB) +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-100]: Local number of parameters: 10G (19074.52MiB) +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Local number of parameters: 10G (19074.52MiB) +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=6|ip-26-0-160-100]: Local number of parameters: 10G (19074.52MiB) +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=7|ip-26-0-160-100]: [After model building] Memory usage: 19074.60MiB. Peak allocated: 19074.67MiB Peak reserved: 20760.00MiB +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=4|ip-26-0-160-100]: [After model building] Memory usage: 19074.60MiB. Peak allocated: 19074.67MiB Peak reserved: 20760.00MiB +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=5|ip-26-0-160-100]: [After model building] Memory usage: 19074.60MiB. Peak allocated: 19074.67MiB Peak reserved: 20760.00MiB +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-100]: [After model building] Memory usage: 19074.60MiB. Peak allocated: 19074.67MiB Peak reserved: 20760.00MiB +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-100]: [After model building] Memory usage: 19074.60MiB. Peak allocated: 19074.67MiB Peak reserved: 20760.00MiB +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-100]: [After model building] Memory usage: 19074.60MiB. Peak allocated: 19074.67MiB Peak reserved: 20760.00MiB +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: [After model building] Memory usage: 19074.60MiB. Peak allocated: 19074.67MiB Peak reserved: 20760.00MiB +12/14/2024 03:59:10 [INFO|DP=0|PP=0|TP=6|ip-26-0-160-100]: [After model building] Memory usage: 19074.60MiB. Peak allocated: 19074.67MiB Peak reserved: 20760.00MiB +12/14/2024 03:59:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: No checkpoint path provided. +12/14/2024 03:59:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Parametrizing model parameters using StandardParametrizator +12/14/2024 03:59:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: [Optimizer Building] Using LearningRateForSP as learning rate +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + optimizer = optimizer_builder(named_param_groups) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.62 GiB is free. Including non-PyTorch memory, this process has 46.70 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.31 GiB is free. Including non-PyTorch memory, this process has 47.01 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.31 GiB is free. Including non-PyTorch memory, this process has 47.01 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + optimizer = optimizer_builder(named_param_groups) + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.54 GiB is free. Including non-PyTorch memory, this process has 46.78 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.62 GiB is free. Including non-PyTorch memory, this process has 46.70 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.54 GiB is free. Including non-PyTorch memory, this process has 46.77 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 33.13 GiB is free. Including non-PyTorch memory, this process has 46.19 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.31 GiB is free. Including non-PyTorch memory, this process has 47.01 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.33 GiB is free. Including non-PyTorch memory, this process has 46.98 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + result = OptimizerFromGradientAccumulator( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.39 GiB is free. Including non-PyTorch memory, this process has 46.93 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 33.33 GiB is free. Including non-PyTorch memory, this process has 45.99 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 33.16 GiB is free. Including non-PyTorch memory, this process has 46.16 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(named_param_groups) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.31 GiB is free. Including non-PyTorch memory, this process has 47.01 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.70 GiB is free. Including non-PyTorch memory, this process has 46.62 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.70 GiB is free. Including non-PyTorch memory, this process has 46.62 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 33.01 GiB is free. Including non-PyTorch memory, this process has 46.30 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.46 GiB is free. Including non-PyTorch memory, this process has 46.86 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.62 GiB is free. Including non-PyTorch memory, this process has 46.70 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.31 GiB is free. Including non-PyTorch memory, this process has 47.01 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.33 GiB is free. Including non-PyTorch memory, this process has 46.98 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.62 GiB is free. Including non-PyTorch memory, this process has 46.70 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.46 GiB is free. Including non-PyTorch memory, this process has 46.86 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.54 GiB is free. Including non-PyTorch memory, this process has 46.77 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + Traceback (most recent call last): +contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + optimizer = optimizer_builder(named_param_groups) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(named_param_groups) + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + result = OptimizerFromGradientAccumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + result = OptimizerFromGradientAccumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + optimizer = optimizer_builder(named_param_groups) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.70 GiB is free. Including non-PyTorch memory, this process has 46.62 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.23 GiB is free. Including non-PyTorch memory, this process has 47.09 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.31 GiB is free. Including non-PyTorch memory, this process has 47.01 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + torch.cudaself.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator(. +OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.70 GiB is free. Including non-PyTorch memory, this process has 46.62 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.23 GiB is free. Including non-PyTorch memory, this process has 47.09 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 33.13 GiB is free. Including non-PyTorch memory, this process has 46.19 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 33.01 GiB is free. Including non-PyTorch memory, this process has 46.30 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.39 GiB is free. Including non-PyTorch memory, this process has 46.93 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.46 GiB is free. Including non-PyTorch memory, this process has 46.86 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.39 GiB is free. Including non-PyTorch memory, this process has 46.93 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.33 GiB is free. Including non-PyTorch memory, this process has 46.98 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.31 GiB is free. Including non-PyTorch memory, this process has 47.01 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.41 GiB is free. Including non-PyTorch memory, this process has 46.91 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(named_param_groups) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.38 GiB is free. Including non-PyTorch memory, this process has 46.94 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + optimizer = optimizer_builder(named_param_groups) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.46 GiB is free. Including non-PyTorch memory, this process has 46.86 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.62 GiB is free. Including non-PyTorch memory, this process has 46.70 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.39 GiB is free. Including non-PyTorch memory, this process has 46.93 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + Traceback (most recent call last): + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + trainer = DistributedTrainer(config_file) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.54 GiB is free. Including non-PyTorch memory, this process has 46.78 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + trainer = DistributedTrainer(config_file) + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( +Traceback (most recent call last): +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 33.13 GiB is free. Including non-PyTorch memory, this process has 46.19 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + optimizer = optimizer_builder(named_param_groups) + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + result = OptimizerFromGradientAccumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + trainer = DistributedTrainer(config_file) + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 33.13 GiB is free. Including non-PyTorch memory, this process has 46.19 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) +Traceback (most recent call last): + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + result = OptimizerFromGradientAccumulator( + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.38 GiB is free. Including non-PyTorch memory, this process has 46.94 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + result = OptimizerFromGradientAccumulator( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + optimizer = optimizer_builder(named_param_groups) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + optimizer = optimizer_builder(named_param_groups) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.49 GiB is free. Including non-PyTorch memory, this process has 46.83 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.41 GiB is free. Including non-PyTorch memory, this process has 46.91 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.54 GiB is free. Including non-PyTorch memory, this process has 46.77 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + result = OptimizerFromGradientAccumulator( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 33.13 GiB is free. Including non-PyTorch memory, this process has 46.19 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items())gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.54 GiB is free. Including non-PyTorch memory, this process has 46.78 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator(self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.62 GiB is free. Including non-PyTorch memory, this process has 46.70 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 3 has a total capacty of 79.33 GiB of which 32.23 GiB is free. Including non-PyTorch memory, this process has 47.09 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.54 GiB is free. Including non-PyTorch memory, this process has 46.78 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + result = OptimizerFromGradientAccumulator( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 33.13 GiB is free. Including non-PyTorch memory, this process has 46.19 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 32.62 GiB is free. Including non-PyTorch memory, this process has 46.70 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + trainer = DistributedTrainer(config_file) + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.46 GiB is free. Including non-PyTorch memory, this process has 46.86 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.41 GiB is free. Including non-PyTorch memory, this process has 46.91 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + result = OptimizerFromGradientAccumulator( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + trainer = DistributedTrainer(config_file) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.41 GiB is free. Including non-PyTorch memory, this process has 46.91 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + trainer = DistributedTrainer(config_file) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.38 GiB is free. Including non-PyTorch memory, this process has 46.94 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + trainer = DistributedTrainer(config_file) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + result = OptimizerFromGradientAccumulator( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.46 GiB is free. Including non-PyTorch memory, this process has 46.86 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 32.46 GiB is free. Including non-PyTorch memory, this process has 46.86 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + result = OptimizerFromGradientAccumulator( + optimizer = optimizer_builder(named_param_groups) + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.49 GiB is free. Including non-PyTorch memory, this process has 46.83 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 6 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.41 GiB is free. Including non-PyTorch memory, this process has 46.91 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + optimizer = optimizer_builder(named_param_groups) + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + result = OptimizerFromGradientAccumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 5 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.49 GiB is free. Including non-PyTorch memory, this process has 46.83 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + result = OptimizerFromGradientAccumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 7 has a total capacty of 79.33 GiB of which 33.33 GiB is free. Including non-PyTorch memory, this process has 45.99 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.41 GiB is free. Including non-PyTorch memory, this process has 46.91 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.51 GiB is free. Including non-PyTorch memory, this process has 46.81 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.43 GiB is free. Including non-PyTorch memory, this process has 46.89 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + result = OptimizerFromGradientAccumulator( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 32.35 GiB is free. Including non-PyTorch memory, this process has 46.97 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 33.12 GiB is free. Including non-PyTorch memory, this process has 46.20 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( +contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 0 has a total capacty of 79.33 GiB of which 33.16 GiB is free. Including non-PyTorch memory, this process has 46.16 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.49 GiB is free. Including non-PyTorch memory, this process has 46.83 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 1 has a total capacty of 79.33 GiB of which 33.13 GiB is free. Including non-PyTorch memory, this process has 46.19 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 32.41 GiB is free. Including non-PyTorch memory, this process has 46.91 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 4 has a total capacty of 79.33 GiB of which 33.12 GiB is free. Including non-PyTorch memory, this process has 46.20 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 37.26 GiB. GPU 2 has a total capacty of 79.33 GiB of which 33.13 GiB is free. Including non-PyTorch memory, this process has 46.19 GiB memory in use. Of the allocated memory 37.25 GiB is allocated by PyTorch, and 662.88 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[2024-12-14 03:59:30,345] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31338 closing signal SIGTERM +[2024-12-14 03:59:30,345] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31339 closing signal SIGTERM +[2024-12-14 03:59:30,345] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31340 closing signal SIGTERM +[2024-12-14 03:59:30,345] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31341 closing signal SIGTERM +[2024-12-14 03:59:30,345] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31342 closing signal SIGTERM +[2024-12-14 03:59:30,345] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31344 closing signal SIGTERM +[2024-12-14 03:59:30,345] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31345 closing signal SIGTERM +[2024-12-14 03:59:33,541] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 5 (pid: 31343) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-14_03:59:30 + host : ip-26-0-160-100.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 31343) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-160-100: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13440897.0 +[2024-12-14 03:59:34,099] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,099] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130990 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,100] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130991 closing signal SIGTERM +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104392 closing signal SIGTERM +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130992 closing signal SIGTERM +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104393 closing signal SIGTERM +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104394 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127623 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75415 closing signal SIGTERM +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104395 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127624 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75416 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127625 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104396 closing signal SIGTERM +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130994 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75417 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127626 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75418 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104397 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68432 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75420 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104398 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127628 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75421 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130995 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68434 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75422 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127629 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 75423 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130996 closing signal SIGTERM +[2024-12-14 03:59:34,099] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130997 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68435 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127630 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68436 closing signal SIGTERM +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98517 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68437 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68438 closing signal SIGTERM +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98518 closing signal SIGTERM +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98519 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68439 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99762 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81059 closing signal SIGTERM +[2024-12-14 03:59:34,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98520 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81060 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99763 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98521 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81061 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98522 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81062 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98523 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 38343 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99764 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81063 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98524 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81064 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99765 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81065 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99766 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 81066 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 38344 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103427 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99767 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 38345 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99618 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99768 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103814 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 38346 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103428 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99619 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99769 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 38347 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103429 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105969 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99620 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103815 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 38348 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98743 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105970 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99621 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103816 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105971 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103817 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103431 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105972 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99622 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103818 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103432 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98744 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105973 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99623 closing signal SIGTERM +[2024-12-14 03:59:34,102] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98952 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103819 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105974 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103820 closing signal SIGTERM +[2024-12-14 03:59:34,102] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103433 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98745 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105975 closing signal SIGTERM +[2024-12-14 03:59:34,102] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98953 closing signal SIGTERM +[2024-12-14 03:59:34,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103821 closing signal SIGTERM +[2024-12-14 03:59:34,102] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103434 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105976 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98746 closing signal SIGTERM +[2024-12-14 03:59:34,102] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98955 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98747 closing signal SIGTERM +[2024-12-14 03:59:34,102] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98956 closing signal SIGTERM +[2024-12-14 03:59:34,102] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98957 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98748 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98749 closing signal SIGTERM +[2024-12-14 03:59:34,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98750 closing signal SIGTERM +[2024-12-14 03:59:34,321] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-164.ec2.internal_103742_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +[2024-12-14 03:59:35,009] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-78.ec2.internal_98445_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent +[2024-12-14 03:59:35,015] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-160-103.ec2.internal_68361_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 98879 got signal: 15 +[2024-12-14 03:59:35,075] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-160-242.ec2.internal_130918_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +[2024-12-14 03:59:35,131] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-213.ec2.internal_105896_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +[2024-12-14 03:59:35,198] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-202.ec2.internal_38270_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +[2024-12-14 03:59:35,221] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-160-225.ec2.internal_127551_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +[2024-12-14 03:59:35,247] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-59.ec2.internal_103354_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +[2024-12-14 03:59:35,253] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-116.ec2.internal_80988_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +[2024-12-14 03:59:35,263] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-103.ec2.internal_99690_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +[2024-12-14 03:59:35,267] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-123.ec2.internal_104320_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +[2024-12-14 03:59:35,287] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-147.ec2.internal_98670_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +[2024-12-14 03:59:35,299] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-131.ec2.internal_75343_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +[2024-12-14 03:59:35,310] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-142.ec2.internal_99545_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-172-252: task 15: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 38270 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 99545 got signal: 15 +srun: error: ip-26-0-165-202: task 10: Exited with exit code 1 +srun: error: ip-26-0-172-142: task 13: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 127551 got signal: 15 +srun: error: ip-26-0-160-225: task 2: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 68361 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 103354 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 98445 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 98670 got signal: 15 +srun: error: ip-26-0-160-103: task 1: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 105896 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 75343 got signal: 15 +srun: error: ip-26-0-165-59: task 7: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ +torch.distributed.elastic.multiprocessing.api.SignalException: Process 104320 got signal: 15 + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 130918 got signal: 15 +srun: error: ip-26-0-161-78: task 4: Exited with exit code 1 +srun: error: ip-26-0-172-147: task 14: Exited with exit code 1 +srun: error: ip-26-0-165-131: task 8: Exited with exit code 1 +srun: error: ip-26-0-165-213: task 11: Exited with exit code 1 +srun: error: ip-26-0-161-123: task 6: Exited with exit code 1 +srun: error: ip-26-0-160-242: task 3: Exited with exit code 1 +[2024-12-14 03:59:39,322] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-164.ec2.internal_103742_0' has failed to send a keep-alive heartbeat to the rendezvous '13440897' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 99690 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 80988 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 103742 got signal: 15 +srun: error: ip-26-0-172-116: task 12: Exited with exit code 1 +srun: error: ip-26-0-165-164: task 9: Exited with exit code 1 +srun: error: ip-26-0-161-103: task 5: Exited with exit code 1 diff --git a/logs/13442145-bench_80G_dp1_tp1_pp8_acc2_mbs128_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/13442145-bench_80G_dp1_tp1_pp8_acc2_mbs128_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..c3d4d21d67f37d88be3c27e701b235bd653d139a --- /dev/null +++ b/logs/13442145-bench_80G_dp1_tp1_pp8_acc2_mbs128_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,805 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-171-56 ++ export NODELIST=ip-26-0-171-56 ++ NODELIST=ip-26-0-171-56 +++ scontrol show hostnames ip-26-0-171-56 +++ head -n1 ++ export MASTER_NODE=ip-26-0-171-56 ++ MASTER_NODE=ip-26-0-171-56 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-43 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 52-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 52-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 52-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 52-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-171-56' +Master node: ip-26-0-171-56 ++ echo 'All nodes: ip-26-0-171-56' +All nodes: ip-26-0-171-56 ++ echo 'World size: 8' +World size: 8 ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13442145 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-171-56:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_80G_dp1_tp1_pp8_acc2_mbs128_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-14 03:10:53,885] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 03:10:53,885] torch.distributed.run: [WARNING] +[2024-12-14 03:10:53,885] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 03:10:53,885] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 03:10:53,885] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +12/14/2024 03:11:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: Measuring inter-GPU and intra-node bandwidth... +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: Bandwidth measurement complete. Time taken: 14.43 seconds +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: Config: +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: Config(general=GeneralArgs(project='debug', +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: run='80G_dp1_tp1_pp8_acc2_mbs128_seq4096_zero0_tpmodeRED_vocab131k', +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: seed=42, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: step=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: consumed_train_samples=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: ignore_sanity_checks=True), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: parallelism=ParallelismArgs(dp=1, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: pp=8, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: tp=1, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: pp_engine=, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: tp_mode=, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: tp_linear_async_communication=True, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: recompute_layer=False, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: tp_recompute_allgather=True, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: expert_parallel_size=1), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: eos_token_id=0, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: hidden_act='silu', +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: hidden_size=8192, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: initializer_range=0.02, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: intermediate_size=28672, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: is_llama_config=True, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: max_position_embeddings=4096, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: num_attention_heads=64, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: num_hidden_layers=80, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: num_key_value_heads=64, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: pad_token_id=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: pretraining_tp=1, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: rms_norm_eps=1e-05, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: rope_scaling=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: rope_theta=10000.0, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: rope_interleaved=False, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: tie_word_embeddings=False, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: use_cache=True, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: vocab_size=131072), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: init_method=RandomInit(std=0.02), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: dtype=torch.bfloat16, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: make_vocab_size_divisible_by=1, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: ddp_bucket_cap_mb=25), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: tokenizer_revision=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: tokenizer_max_length=None), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: checkpoint_interval=10000, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: save_initial_state=False, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: save_final_state=False, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: resume_checkpoint_path=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: checkpoints_path_is_shared_file_system=False), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: logging=LoggingArgs(log_level='info', +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: log_level_replica='info', +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: iteration_step_info_interval=1), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: tokens=TokensArgs(sequence_length=4096, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: train_steps=100, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: micro_batch_size=128, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: batch_accumulation_per_replica=2, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: val_check_interval=100, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: limit_val_batches=0, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: limit_test_batches=0), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: adam_beta1=0.9, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: adam_beta2=0.95, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: torch_adam_is_fused=True, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: name='adamW'), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: zero_stage=0, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: weight_decay=0.01, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: clip_grad=1.0, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: accumulate_grad_in_fp32=True, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: lr_warmup_steps=2, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: lr_warmup_style='linear', +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: lr_decay_style='cosine', +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: lr_decay_steps=13, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: lr_decay_starting_step=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: min_decay_lr=1e-05)), +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: start_training_step=1, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: data=DataArgs(dataset=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: seed=42, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: num_loading_workers=1))], +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: profiler=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: lighteval=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: s3_upload=None) +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: Model Config: +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: LlamaConfig(bos_token_id=0, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: eos_token_id=0, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: hidden_act='silu', +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: hidden_size=8192, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: initializer_range=0.02, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: intermediate_size=28672, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: is_llama_config=True, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: max_position_embeddings=4096, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: num_attention_heads=64, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: num_hidden_layers=80, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: num_key_value_heads=64, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: pad_token_id=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: pretraining_tp=1, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: rms_norm_eps=1e-05, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: rope_scaling=None, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: rope_theta=10000.0, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: rope_interleaved=False, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: tie_word_embeddings=False, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: use_cache=True, +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: vocab_size=131072) +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: Building model.. +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: Initialize RoPE Theta = 10000.0 +12/14/2024 03:11:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/14/2024 03:11:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: Total number of parameters: 80G (152578.52MiB) +12/14/2024 03:11:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: Local number of parameters: 11.8G (22464.34MiB) +12/14/2024 03:11:31 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-56]: Local number of parameters: 9.73G (18560.31MiB) +12/14/2024 03:11:31 [INFO|DP=0|PP=2|TP=0|ip-26-0-171-56]: Local number of parameters: 9.73G (18560.31MiB) +12/14/2024 03:11:31 [INFO|DP=0|PP=4|TP=0|ip-26-0-171-56]: Local number of parameters: 9.73G (18560.31MiB) +12/14/2024 03:11:31 [INFO|DP=0|PP=6|TP=0|ip-26-0-171-56]: Local number of parameters: 9.73G (18560.31MiB) +12/14/2024 03:11:31 [INFO|DP=0|PP=5|TP=0|ip-26-0-171-56]: Local number of parameters: 9.73G (18560.31MiB) +12/14/2024 03:11:31 [INFO|DP=0|PP=7|TP=0|ip-26-0-171-56]: Local number of parameters: 9.83G (18752.30MiB) +12/14/2024 03:11:31 [INFO|DP=0|PP=3|TP=0|ip-26-0-171-56]: Local number of parameters: 9.73G (18560.31MiB) +12/14/2024 03:11:31 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 18560.33MiB. Peak allocated: 18560.33MiB Peak reserved: 19714.00MiB +12/14/2024 03:11:31 [INFO|DP=0|PP=4|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 18560.33MiB. Peak allocated: 18560.33MiB Peak reserved: 19714.00MiB +12/14/2024 03:11:31 [INFO|DP=0|PP=6|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 18560.33MiB. Peak allocated: 18560.33MiB Peak reserved: 19586.00MiB +12/14/2024 03:11:31 [INFO|DP=0|PP=2|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 18560.33MiB. Peak allocated: 18560.33MiB Peak reserved: 19714.00MiB +12/14/2024 03:11:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 22464.36MiB. Peak allocated: 22464.37MiB Peak reserved: 23490.00MiB +12/14/2024 03:11:31 [INFO|DP=0|PP=5|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 18560.33MiB. Peak allocated: 18560.33MiB Peak reserved: 19714.00MiB +12/14/2024 03:11:31 [INFO|DP=0|PP=7|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 18752.31MiB. Peak allocated: 18752.32MiB Peak reserved: 20034.00MiB +12/14/2024 03:11:31 [INFO|DP=0|PP=3|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 18560.33MiB. Peak allocated: 18560.33MiB Peak reserved: 19586.00MiB +12/14/2024 03:11:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: No checkpoint path provided. +12/14/2024 03:11:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: Parametrizing model parameters using StandardParametrizator +12/14/2024 03:11:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-56]: [Optimizer Building] Using LearningRateForSP as learning rate +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.25 GiB. GPU 1 has a total capacty of 79.33 GiB of which 15.99 GiB is free. Including non-PyTorch memory, this process has 63.33 GiB memory in use. Of the allocated memory 54.38 GiB is allocated by PyTorch, and 1.13 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.25 GiB. GPU 5 has a total capacty of 79.33 GiB of which 15.99 GiB is free. Including non-PyTorch memory, this process has 63.33 GiB memory in use. Of the allocated memory 54.38 GiB is allocated by PyTorch, and 1.13 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.25 GiB. GPU 6 has a total capacty of 79.33 GiB of which 16.12 GiB is free. Including non-PyTorch memory, this process has 63.20 GiB memory in use. Of the allocated memory 54.38 GiB is allocated by PyTorch, and 1.00 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.25 GiB. GPU 2 has a total capacty of 79.33 GiB of which 15.99 GiB is free. Including non-PyTorch memory, this process has 63.33 GiB memory in use. Of the allocated memory 54.38 GiB is allocated by PyTorch, and 1.13 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator(gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.63 GiB. GPU 7 has a total capacty of 79.33 GiB of which 16.13 GiB is free. Including non-PyTorch memory, this process has 63.19 GiB memory in use. Of the allocated memory 54.94 GiB is allocated by PyTorch, and 1.13 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 43.88 GiB. GPU 0 has a total capacty of 79.33 GiB of which 4.82 GiB is free. Including non-PyTorch memory, this process has 74.50 GiB memory in use. Of the allocated memory 65.81 GiB is allocated by PyTorch, and 1.00 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator +gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.25 GiB. GPU 4 has a total capacty of 79.33 GiB of which 15.99 GiB is free. Including non-PyTorch memory, this process has 63.33 GiB memory in use. Of the allocated memory 54.38 GiB is allocated by PyTorch, and 1.13 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.25 GiB. GPU 3 has a total capacty of 79.33 GiB of which 16.12 GiB is free. Including non-PyTorch memory, this process has 63.20 GiB memory in use. Of the allocated memory 54.38 GiB is allocated by PyTorch, and 1.00 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[2024-12-14 03:11:44,060] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 89522) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-14_03:11:44 + host : ip-26-0-171-56.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 89523) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-14_03:11:44 + host : ip-26-0-171-56.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 89524) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-14_03:11:44 + host : ip-26-0-171-56.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 89525) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-14_03:11:44 + host : ip-26-0-171-56.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 89526) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-14_03:11:44 + host : ip-26-0-171-56.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 89527) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-14_03:11:44 + host : ip-26-0-171-56.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 89528) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-14_03:11:44 + host : ip-26-0-171-56.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 89529) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-14_03:11:44 + host : ip-26-0-171-56.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 89522) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-171-56: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13442145.0 diff --git a/logs/13442619-bench_8.86G_dp1_tp1_pp8_acc2_mbs128_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/13442619-bench_8.86G_dp1_tp1_pp8_acc2_mbs128_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..b3024af26877ac77c7438ff0eea2987a48812aae --- /dev/null +++ b/logs/13442619-bench_8.86G_dp1_tp1_pp8_acc2_mbs128_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,715 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n x ']' +++++ manpath ++++ [[ ! :/admin/home/nouamane/miniconda/envs/2-1-cu121/man:/admin/home/nouamane/miniconda/envs/2-1-cu121/share/man:/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib::/opt/amazon/openmpi/lib:/opt/amazon/efa/lib;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=:1:/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib::/opt/amazon/openmpi/lib:/opt/amazon/efa/lib +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=:1:/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/admin/home/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/admin/home/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/admin/home/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/admin/home/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /admin/home/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/admin/home/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/admin/home/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ scontrol show hostnames ip-26-0-172-147 ++ export NODELIST=ip-26-0-172-147 ++ NODELIST=ip-26-0-172-147 +++ scontrol show hostnames ip-26-0-172-147 +++ head -n1 ++ export MASTER_NODE=ip-26-0-172-147 ++ MASTER_NODE=ip-26-0-172-147 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-172-147' +Master node: ip-26-0-172-147 ++ echo 'All nodes: ip-26-0-172-147' +All nodes: ip-26-0-172-147 ++ echo 'World size: 8' +World size: 8 ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13442619 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-172-147:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_8.86G_dp1_tp1_pp8_acc2_mbs128_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-14 07:40:18,022] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 07:40:18,022] torch.distributed.run: [WARNING] +[2024-12-14 07:40:18,022] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 07:40:18,022] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 07:40:18,022] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.6+cuda12.1 +12/14/2024 07:40:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Measuring inter-GPU and intra-node bandwidth... +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Bandwidth measurement complete. Time taken: 14.16 seconds +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Config: +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Config(general=GeneralArgs(project='debug', +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: run='8.86G_dp1_tp1_pp8_acc2_mbs128_seq4096_zero0_tpmodeRED_vocab131k', +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: seed=42, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: step=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: consumed_train_samples=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: ignore_sanity_checks=True), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: parallelism=ParallelismArgs(dp=1, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: pp=8, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: tp=1, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: pp_engine=, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: tp_mode=, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: tp_linear_async_communication=True, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: recompute_layer=False, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: tp_recompute_allgather=True, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: expert_parallel_size=1), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: eos_token_id=0, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: hidden_act='silu', +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: hidden_size=4096, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: initializer_range=0.02, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: intermediate_size=14336, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: is_llama_config=True, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: max_position_embeddings=4096, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: num_attention_heads=32, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: num_hidden_layers=32, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: num_key_value_heads=32, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: pad_token_id=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: pretraining_tp=1, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: rms_norm_eps=1e-05, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: rope_scaling=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: rope_theta=10000.0, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: rope_interleaved=False, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: tie_word_embeddings=False, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: use_cache=True, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: vocab_size=131072), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: init_method=RandomInit(std=0.02), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: dtype=torch.bfloat16, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: make_vocab_size_divisible_by=1, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: ddp_bucket_cap_mb=25), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: tokenizer_revision=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: tokenizer_max_length=None), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: checkpoint_interval=10000, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: save_initial_state=False, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: save_final_state=False, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: resume_checkpoint_path=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: checkpoints_path_is_shared_file_system=False), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: logging=LoggingArgs(log_level='info', +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: log_level_replica='info', +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: iteration_step_info_interval=1), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: tokens=TokensArgs(sequence_length=4096, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: train_steps=100, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: micro_batch_size=128, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: batch_accumulation_per_replica=2, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: val_check_interval=100, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: limit_val_batches=0, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: limit_test_batches=0), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: adam_beta1=0.9, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: adam_beta2=0.95, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: torch_adam_is_fused=True, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: name='adamW'), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: zero_stage=0, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: weight_decay=0.01, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: clip_grad=1.0, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: accumulate_grad_in_fp32=True, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: lr_warmup_steps=2, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: lr_warmup_style='linear', +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: lr_decay_style='cosine', +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: lr_decay_steps=13, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: lr_decay_starting_step=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: min_decay_lr=1e-05)), +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: start_training_step=1, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: data=DataArgs(dataset=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: seed=42, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: num_loading_workers=1))], +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: profiler=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: lighteval=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: s3_upload=None) +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Model Config: +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: LlamaConfig(bos_token_id=0, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: eos_token_id=0, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: hidden_act='silu', +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: hidden_size=4096, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: initializer_range=0.02, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: intermediate_size=14336, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: is_llama_config=True, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: max_position_embeddings=4096, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: num_attention_heads=32, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: num_hidden_layers=32, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: num_key_value_heads=32, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: pad_token_id=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: pretraining_tp=1, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: rms_norm_eps=1e-05, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: rope_scaling=None, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: rope_theta=10000.0, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: rope_interleaved=False, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: tie_word_embeddings=False, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: use_cache=True, +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: vocab_size=131072) +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Building model.. +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Initialize RoPE Theta = 10000.0 +12/14/2024 07:40:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Setting PP block ranks... +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Total number of parameters: 8.86G (16896.51MiB) +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Local number of parameters: 1.75G (3344.08MiB) +12/14/2024 07:40:52 [INFO|DP=0|PP=7|TP=0|ip-26-0-172-147]: Local number of parameters: 1.02G (1952.04MiB) +12/14/2024 07:40:52 [INFO|DP=0|PP=6|TP=0|ip-26-0-172-147]: Local number of parameters: 973M (1856.06MiB) +12/14/2024 07:40:52 [INFO|DP=0|PP=1|TP=0|ip-26-0-172-147]: Local number of parameters: 973M (1856.06MiB) +12/14/2024 07:40:52 [INFO|DP=0|PP=5|TP=0|ip-26-0-172-147]: Local number of parameters: 973M (1856.06MiB) +12/14/2024 07:40:52 [INFO|DP=0|PP=2|TP=0|ip-26-0-172-147]: Local number of parameters: 973M (1856.06MiB) +12/14/2024 07:40:52 [INFO|DP=0|PP=4|TP=0|ip-26-0-172-147]: Local number of parameters: 973M (1856.06MiB) +12/14/2024 07:40:52 [INFO|DP=0|PP=3|TP=0|ip-26-0-172-147]: Local number of parameters: 1.22G (2320.08MiB) +12/14/2024 07:40:52 [INFO|DP=0|PP=7|TP=0|ip-26-0-172-147]: [After model building] Memory usage: 1952.05MiB. Peak allocated: 5632.00MiB Peak reserved: 11778.00MiB +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: [After model building] Memory usage: 3344.09MiB. Peak allocated: 5632.00MiB Peak reserved: 11778.00MiB +12/14/2024 07:40:52 [INFO|DP=0|PP=6|TP=0|ip-26-0-172-147]: [After model building] Memory usage: 1856.08MiB. Peak allocated: 5632.00MiB Peak reserved: 11778.00MiB +12/14/2024 07:40:52 [INFO|DP=0|PP=3|TP=0|ip-26-0-172-147]: [After model building] Memory usage: 2320.09MiB. Peak allocated: 5632.00MiB Peak reserved: 11778.00MiB +12/14/2024 07:40:52 [INFO|DP=0|PP=1|TP=0|ip-26-0-172-147]: [After model building] Memory usage: 1856.08MiB. Peak allocated: 5632.00MiB Peak reserved: 11778.00MiB +12/14/2024 07:40:52 [INFO|DP=0|PP=2|TP=0|ip-26-0-172-147]: [After model building] Memory usage: 1856.08MiB. Peak allocated: 5632.00MiB Peak reserved: 11778.00MiB +12/14/2024 07:40:52 [INFO|DP=0|PP=5|TP=0|ip-26-0-172-147]: [After model building] Memory usage: 1856.08MiB. Peak allocated: 5632.00MiB Peak reserved: 11778.00MiB +12/14/2024 07:40:52 [INFO|DP=0|PP=4|TP=0|ip-26-0-172-147]: [After model building] Memory usage: 1856.08MiB. Peak allocated: 5632.00MiB Peak reserved: 11778.00MiB +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: No checkpoint path provided. +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Parametrizing model parameters using StandardParametrizator +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: [Optimizer Building] Using LearningRateForSP as learning rate +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Using dummy data generator +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: [Training Plan] There are 1 training stages +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: [Stage Stable Training Stage] start from step 1 +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: +12/14/2024 07:40:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: [Start training] datetime: 2024-12-14 07:40:52.981523 | mbs: 128 | grad_accum: 2 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 243, in train_batch_iter + self.nb_microbatches >= pg.size() - 1 +AssertionError: Number of microbatches (2) must be at least PP_SIZE-1=7 when using the OneForwardOneBackwardPipelineEngine +12/14/2024 07:40:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/14/2024 07:40:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-147]: Memory usage: 16720.40MiB. Peak allocated 16720.40MiB. Peak reserved: 25158.00MiB +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 243, in train_batch_iter + self.nb_microbatches >= pg.size() - 1 +AssertionError: Number of microbatches (2) must be at least PP_SIZE-1=7 when using the OneForwardOneBackwardPipelineEngine +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 243, in train_batch_iter + self.nb_microbatches >= pg.size() - 1 +AssertionError: Number of microbatches (2) must be at least PP_SIZE-1=7 when using the OneForwardOneBackwardPipelineEngine +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 243, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + self.nb_microbatches >= pg.size() - 1 +AssertionError: Number of microbatches (2) must be at least PP_SIZE-1=7 when using the OneForwardOneBackwardPipelineEngine + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 243, in train_batch_iter + self.nb_microbatches >= pg.size() - 1 +AssertionError: Number of microbatches (2) must be at least PP_SIZE-1=7 when using the OneForwardOneBackwardPipelineEngine +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 243, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + self.nb_microbatches >= pg.size() - 1 +AssertionError: Number of microbatches (2) must be at least PP_SIZE-1=7 when using the OneForwardOneBackwardPipelineEngine + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 243, in train_batch_iter + self.nb_microbatches >= pg.size() - 1 +AssertionError: Number of microbatches (2) must be at least PP_SIZE-1=7 when using the OneForwardOneBackwardPipelineEngine +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 243, in train_batch_iter + self.nb_microbatches >= pg.size() - 1 +AssertionError: Number of microbatches (2) must be at least PP_SIZE-1=7 when using the OneForwardOneBackwardPipelineEngine +[2024-12-14 07:41:03,141] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52539 closing signal SIGTERM +[2024-12-14 07:41:03,141] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52540 closing signal SIGTERM +[2024-12-14 07:41:03,141] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52541 closing signal SIGTERM +[2024-12-14 07:41:03,141] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52543 closing signal SIGTERM +[2024-12-14 07:41:03,141] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52544 closing signal SIGTERM +[2024-12-14 07:41:03,141] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52546 closing signal SIGTERM +[2024-12-14 07:41:04,908] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 3 (pid: 52542) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-14_07:41:03 + host : ip-26-0-172-147.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 52545) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-14_07:41:03 + host : ip-26-0-172-147.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 52542) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-172-147: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13442619.0 diff --git a/logs/13443083-bench_80G_dp1_tp2_pp4_acc4_mbs64_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/13443083-bench_80G_dp1_tp2_pp4_acc4_mbs64_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..8e681d2921da3caa71f2a6b90c2a302b1a785d23 --- /dev/null +++ b/logs/13443083-bench_80G_dp1_tp2_pp4_acc4_mbs64_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,764 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames ip-26-0-165-164 ++ export NODELIST=ip-26-0-165-164 ++ NODELIST=ip-26-0-165-164 +++ scontrol show hostnames ip-26-0-165-164 +++ head -n1 ++ export MASTER_NODE=ip-26-0-165-164 ++ MASTER_NODE=ip-26-0-165-164 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-165-164' +Master node: ip-26-0-165-164 ++ echo 'All nodes: ip-26-0-165-164' +All nodes: ip-26-0-165-164 ++ echo 'World size: 8' +World size: 8 ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13443083 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-165-164:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_80G_dp1_tp2_pp4_acc4_mbs64_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-14 09:20:29,769] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 09:20:29,769] torch.distributed.run: [WARNING] +[2024-12-14 09:20:29,769] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 09:20:29,769] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 09:20:29,769] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +12/14/2024 09:20:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: Measuring inter-GPU and intra-node bandwidth... +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: Bandwidth measurement complete. Time taken: 14.03 seconds +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: Config: +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: Config(general=GeneralArgs(project='debug', +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: run='80G_dp1_tp2_pp4_acc4_mbs64_seq4096_zero0_tpmodeRED_vocab131k', +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: seed=42, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: step=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: consumed_train_samples=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: ignore_sanity_checks=True), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: parallelism=ParallelismArgs(dp=1, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: pp=4, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: tp=2, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: pp_engine=, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: tp_mode=, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: tp_linear_async_communication=True, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: recompute_layer=False, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: tp_recompute_allgather=True, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: expert_parallel_size=1), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: eos_token_id=0, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: hidden_act='silu', +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: hidden_size=8192, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: initializer_range=0.02, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: intermediate_size=28672, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: is_llama_config=True, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: max_position_embeddings=4096, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: num_attention_heads=64, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: num_hidden_layers=80, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: num_key_value_heads=64, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: pad_token_id=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: pretraining_tp=1, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: rms_norm_eps=1e-05, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: rope_scaling=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: rope_theta=10000.0, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: rope_interleaved=False, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: tie_word_embeddings=False, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: use_cache=True, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: vocab_size=131072), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: init_method=RandomInit(std=0.02), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: dtype=torch.bfloat16, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: make_vocab_size_divisible_by=1, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: ddp_bucket_cap_mb=25), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: tokenizer_revision=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: tokenizer_max_length=None), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: checkpoint_interval=10000, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: save_initial_state=False, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: save_final_state=False, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: resume_checkpoint_path=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: checkpoints_path_is_shared_file_system=False), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: logging=LoggingArgs(log_level='info', +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: log_level_replica='info', +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: iteration_step_info_interval=1), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: tokens=TokensArgs(sequence_length=4096, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: train_steps=100, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: micro_batch_size=64, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: batch_accumulation_per_replica=4, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: val_check_interval=100, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: limit_val_batches=0, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: limit_test_batches=0), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: adam_beta1=0.9, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: adam_beta2=0.95, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: torch_adam_is_fused=True, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: name='adamW'), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: zero_stage=0, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: weight_decay=0.01, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: clip_grad=1.0, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: accumulate_grad_in_fp32=True, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: lr_warmup_steps=2, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: lr_warmup_style='linear', +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: lr_decay_style='cosine', +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: lr_decay_steps=13, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: lr_decay_starting_step=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: min_decay_lr=1e-05)), +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: start_training_step=1, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: data=DataArgs(dataset=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: seed=42, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: num_loading_workers=1))], +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: profiler=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: lighteval=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: s3_upload=None) +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: Model Config: +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: LlamaConfig(bos_token_id=0, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: eos_token_id=0, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: hidden_act='silu', +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: hidden_size=8192, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: initializer_range=0.02, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: intermediate_size=28672, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: is_llama_config=True, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: max_position_embeddings=4096, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: num_attention_heads=64, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: num_hidden_layers=80, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: num_key_value_heads=64, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: pad_token_id=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: pretraining_tp=1, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: rms_norm_eps=1e-05, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: rope_scaling=None, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: rope_theta=10000.0, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: rope_interleaved=False, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: tie_word_embeddings=False, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: use_cache=True, +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: vocab_size=131072) +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: Building model.. +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: Initialize RoPE Theta = 10000.0 +12/14/2024 09:21:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/14/2024 09:21:14 [INFO|DP=0|PP=3|TP=1|ip-26-0-165-164]: Local number of parameters: 9.78G (18656.61MiB) +12/14/2024 09:21:14 [INFO|DP=0|PP=2|TP=1|ip-26-0-165-164]: Local number of parameters: 9.73G (18560.62MiB) +12/14/2024 09:21:14 [INFO|DP=0|PP=2|TP=0|ip-26-0-165-164]: Local number of parameters: 9.73G (18560.62MiB) +12/14/2024 09:21:14 [INFO|DP=0|PP=0|TP=1|ip-26-0-165-164]: Local number of parameters: 10.8G (20512.66MiB) +12/14/2024 09:21:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: Total number of parameters: 80G (152581.03MiB) +12/14/2024 09:21:14 [INFO|DP=0|PP=1|TP=0|ip-26-0-165-164]: Local number of parameters: 9.73G (18560.62MiB) +12/14/2024 09:21:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: Local number of parameters: 10.8G (20512.66MiB) +12/14/2024 09:21:14 [INFO|DP=0|PP=1|TP=1|ip-26-0-165-164]: Local number of parameters: 9.73G (18560.62MiB) +12/14/2024 09:21:14 [INFO|DP=0|PP=3|TP=0|ip-26-0-165-164]: Local number of parameters: 9.78G (18656.61MiB) +12/14/2024 09:21:14 [INFO|DP=0|PP=2|TP=1|ip-26-0-165-164]: [After model building] Memory usage: 18560.65MiB. Peak allocated: 18560.66MiB Peak reserved: 19714.00MiB +12/14/2024 09:21:14 [INFO|DP=0|PP=2|TP=0|ip-26-0-165-164]: [After model building] Memory usage: 18560.65MiB. Peak allocated: 18560.66MiB Peak reserved: 19714.00MiB +12/14/2024 09:21:14 [INFO|DP=0|PP=1|TP=0|ip-26-0-165-164]: [After model building] Memory usage: 18560.65MiB. Peak allocated: 18560.66MiB Peak reserved: 20034.00MiB +12/14/2024 09:21:14 [INFO|DP=0|PP=1|TP=1|ip-26-0-165-164]: [After model building] Memory usage: 18560.65MiB. Peak allocated: 18560.66MiB Peak reserved: 20034.00MiB +12/14/2024 09:21:14 [INFO|DP=0|PP=3|TP=1|ip-26-0-165-164]: [After model building] Memory usage: 18656.64MiB. Peak allocated: 18656.64MiB Peak reserved: 19874.00MiB +12/14/2024 09:21:14 [INFO|DP=0|PP=0|TP=1|ip-26-0-165-164]: [After model building] Memory usage: 20512.69MiB. Peak allocated: 20512.69MiB Peak reserved: 21602.00MiB +12/14/2024 09:21:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: [After model building] Memory usage: 20512.69MiB. Peak allocated: 20512.69MiB Peak reserved: 21602.00MiB +12/14/2024 09:21:14 [INFO|DP=0|PP=3|TP=0|ip-26-0-165-164]: [After model building] Memory usage: 18656.64MiB. Peak allocated: 18656.64MiB Peak reserved: 19874.00MiB +12/14/2024 09:21:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: No checkpoint path provided. +12/14/2024 09:21:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: Parametrizing model parameters using StandardParametrizator +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.25 GiB. GPU 5 has a total capacty of 79.33 GiB of which 13.64 GiB is free. Including non-PyTorch memory, this process has 65.67 GiB memory in use. Of the allocated memory 54.38 GiB is allocated by PyTorch, and 1.13 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +12/14/2024 09:21:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-165-164]: [Optimizer Building] Using LearningRateForSP as learning rate +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.25 GiB. GPU 2 has a total capacty of 79.33 GiB of which 13.33 GiB is free. Including non-PyTorch memory, this process has 65.99 GiB memory in use. Of the allocated memory 54.38 GiB is allocated by PyTorch, and 1.44 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.25 GiB. GPU 4 has a total capacty of 79.33 GiB of which 13.64 GiB is free. Including non-PyTorch memory, this process has 65.67 GiB memory in use. Of the allocated memory 54.38 GiB is allocated by PyTorch, and 1.13 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.44 GiB. GPU 7 has a total capacty of 79.33 GiB of which 14.24 GiB is free. Including non-PyTorch memory, this process has 65.08 GiB memory in use. Of the allocated memory 54.66 GiB is allocated by PyTorch, and 1.19 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 40.06 GiB. GPU 1 has a total capacty of 79.33 GiB of which 8.04 GiB is free. Including non-PyTorch memory, this process has 71.28 GiB memory in use. Of the allocated memory 60.10 GiB is allocated by PyTorch, and 1.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.44 GiB. GPU 6 has a total capacty of 79.33 GiB of which 13.54 GiB is free. Including non-PyTorch memory, this process has 65.78 GiB memory in use. Of the allocated memory 54.66 GiB is allocated by PyTorch, and 1.19 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + trainer = DistributedTrainer(config_file)big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 183, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 36.25 GiB. GPU 3 has a total capacty of 79.33 GiB of which 13.33 GiB is free. Including non-PyTorch memory, this process has 65.99 GiB memory in use. Of the allocated memory 54.38 GiB is allocated by PyTorch, and 1.44 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 402, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(named_param_groups) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 95, in __init__ + big_flat_buffer = torch.empty(length, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 40.06 GiB. GPU 0 has a total capacty of 79.33 GiB of which 8.18 GiB is free. Including non-PyTorch memory, this process has 71.14 GiB memory in use. Of the allocated memory 60.10 GiB is allocated by PyTorch, and 1.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[2024-12-14 09:21:24,991] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19640 closing signal SIGTERM +[2024-12-14 09:21:24,992] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19641 closing signal SIGTERM +[2024-12-14 09:21:24,992] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19642 closing signal SIGTERM +[2024-12-14 09:21:24,992] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19643 closing signal SIGTERM +[2024-12-14 09:21:24,992] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19644 closing signal SIGTERM +[2024-12-14 09:21:24,992] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19645 closing signal SIGTERM +[2024-12-14 09:21:24,992] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19646 closing signal SIGTERM +[2024-12-14 09:21:28,662] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 7 (pid: 19647) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-14_09:21:24 + host : ip-26-0-165-164.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 19647) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-165-164: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13443083.0 diff --git a/logs/13446329-bench_1.34G_dp1_tp4_pp8_acc32_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/13446329-bench_1.34G_dp1_tp4_pp8_acc32_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..1483e13510f7a37f8cc275fdda797ad8425c5f78 --- /dev/null +++ b/logs/13446329-bench_1.34G_dp1_tp4_pp8_acc32_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,776 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n x ']' +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/slurm/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' +++ eval 'PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_SHLVL=1 ++++ CONDA_SHLVL=1 ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=2 +++ CONDA_SHLVL=2 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ scontrol show hostnames 'ip-26-0-161-123,ip-26-0-162-14,ip-26-0-168-[120,238]' ++ export 'NODELIST=ip-26-0-161-123 +ip-26-0-162-14 +ip-26-0-168-120 +ip-26-0-168-238' ++ NODELIST='ip-26-0-161-123 +ip-26-0-162-14 +ip-26-0-168-120 +ip-26-0-168-238' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-161-123,ip-26-0-162-14,ip-26-0-168-[120,238]' ++ export MASTER_NODE=ip-26-0-161-123 ++ MASTER_NODE=ip-26-0-161-123 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-161-123' +Master node: ip-26-0-161-123 ++ echo 'All nodes: ip-26-0-161-123 +ip-26-0-162-14 +ip-26-0-168-120 +ip-26-0-168-238' +All nodes: ip-26-0-161-123 +ip-26-0-162-14 +ip-26-0-168-120 +ip-26-0-168-238 ++ echo 'World size: 32' +World size: 32 ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13446329 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-161-123:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.34G_dp1_tp4_pp8_acc32_mbs8_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-14 23:37:48,655] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 23:37:48,654] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 23:37:48,658] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 23:37:48,660] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-14 23:37:48,655] torch.distributed.run: [WARNING] +[2024-12-14 23:37:48,655] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 23:37:48,655] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 23:37:48,655] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 23:37:48,655] torch.distributed.run: [WARNING] +[2024-12-14 23:37:48,655] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 23:37:48,655] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 23:37:48,655] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 23:37:48,659] torch.distributed.run: [WARNING] +[2024-12-14 23:37:48,659] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 23:37:48,659] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 23:37:48,659] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 23:37:48,660] torch.distributed.run: [WARNING] +[2024-12-14 23:37:48,660] torch.distributed.run: [WARNING] ***************************************** +[2024-12-14 23:37:48,660] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-14 23:37:48,660] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +12/14/2024 23:38:12 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Measuring inter-GPU and intra-node bandwidth... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Bandwidth measurement complete. Time taken: 16.96 seconds +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Config: +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Config(general=GeneralArgs(project='debug', +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: run='1.34G_dp1_tp4_pp8_acc32_mbs8_seq4096_zero0_tpmodeRED_vocab131k', +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: seed=42, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: step=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: consumed_train_samples=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: ignore_sanity_checks=True), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: parallelism=ParallelismArgs(dp=1, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: pp=8, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: tp=4, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: pp_engine=, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: tp_mode=, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: tp_linear_async_communication=True, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: recompute_layer=False, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: tp_recompute_allgather=True, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: expert_parallel_size=1), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: eos_token_id=0, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: hidden_act='silu', +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: hidden_size=2048, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: initializer_range=0.02, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: intermediate_size=8192, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: is_llama_config=True, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: max_position_embeddings=4096, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: num_attention_heads=32, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: num_hidden_layers=16, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: num_key_value_heads=32, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: pad_token_id=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: pretraining_tp=1, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: rms_norm_eps=1e-05, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: rope_scaling=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: rope_theta=10000.0, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: rope_interleaved=False, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: tie_word_embeddings=True, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: use_cache=True, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: vocab_size=131072), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: init_method=RandomInit(std=0.02), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: dtype=torch.bfloat16, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: make_vocab_size_divisible_by=1, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: ddp_bucket_cap_mb=25), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: tokenizer_revision=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: tokenizer_max_length=None), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: checkpoint_interval=10000, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: save_initial_state=False, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: save_final_state=False, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: resume_checkpoint_path=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: checkpoints_path_is_shared_file_system=False), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: logging=LoggingArgs(log_level='info', +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: log_level_replica='info', +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: iteration_step_info_interval=1), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: tokens=TokensArgs(sequence_length=4096, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: train_steps=100, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: micro_batch_size=8, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: batch_accumulation_per_replica=32, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: val_check_interval=100, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: limit_val_batches=0, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: limit_test_batches=0), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: adam_beta1=0.9, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: adam_beta2=0.95, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: torch_adam_is_fused=True, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: name='adamW'), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: zero_stage=0, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: weight_decay=0.01, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: clip_grad=1.0, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: accumulate_grad_in_fp32=True, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: lr_warmup_steps=2, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: lr_warmup_style='linear', +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: lr_decay_style='cosine', +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: lr_decay_steps=13, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: lr_decay_starting_step=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: min_decay_lr=1e-05)), +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: start_training_step=1, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: data=DataArgs(dataset=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: seed=42, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: num_loading_workers=1))], +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: profiler=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: lighteval=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: s3_upload=None) +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Model Config: +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: LlamaConfig(bos_token_id=0, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: eos_token_id=0, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: hidden_act='silu', +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: hidden_size=2048, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: initializer_range=0.02, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: intermediate_size=8192, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: is_llama_config=True, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: max_position_embeddings=4096, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: num_attention_heads=32, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: num_hidden_layers=16, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: num_key_value_heads=32, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: pad_token_id=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: pretraining_tp=1, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: rms_norm_eps=1e-05, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: rope_scaling=None, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: rope_theta=10000.0, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: rope_interleaved=False, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: tie_word_embeddings=True, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: use_cache=True, +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: vocab_size=131072) +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Building model.. +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Initialize RoPE Theta = 10000.0 +12/14/2024 23:38:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/14/2024 23:38:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Total number of parameters: 1.61G (3072.52MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Local number of parameters: 117M (224.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=1|TP=0|ip-26-0-161-123]: Local number of parameters: 50.3M (96.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: [After model building] Memory usage: 224.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=1|TP=0|ip-26-0-161-123]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: No checkpoint path provided. +12/14/2024 23:38:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Parametrizing model parameters using StandardParametrizator +12/14/2024 23:38:36 [INFO|DP=0|PP=1|TP=3|ip-26-0-161-123]: Local number of parameters: 50.3M (96.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=0|TP=3|ip-26-0-161-123]: Local number of parameters: 117M (224.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=1|TP=3|ip-26-0-161-123]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=0|TP=3|ip-26-0-161-123]: [After model building] Memory usage: 224.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=1|TP=2|ip-26-0-161-123]: Local number of parameters: 50.3M (96.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=0|TP=2|ip-26-0-161-123]: Local number of parameters: 117M (224.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=4|TP=0|ip-26-0-168-120]: Local number of parameters: 33.6M (64.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=5|TP=0|ip-26-0-168-120]: Local number of parameters: 50.3M (96.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=1|TP=2|ip-26-0-161-123]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=4|TP=0|ip-26-0-168-120]: [After model building] Memory usage: 64.03MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=0|TP=2|ip-26-0-161-123]: [After model building] Memory usage: 224.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=5|TP=0|ip-26-0-168-120]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=0|TP=1|ip-26-0-161-123]: Local number of parameters: 117M (224.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=1|TP=1|ip-26-0-161-123]: Local number of parameters: 50.3M (96.02MiB) +NCCL version 2.18.5+cuda12.2 +12/14/2024 23:38:36 [INFO|DP=0|PP=0|TP=1|ip-26-0-161-123]: [After model building] Memory usage: 224.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=1|TP=1|ip-26-0-161-123]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=5|TP=3|ip-26-0-168-120]: Local number of parameters: 50.3M (96.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=7|TP=3|ip-26-0-168-238]: Local number of parameters: 0 (0.00MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=3|TP=3|ip-26-0-162-14]: Local number of parameters: 50.3M (96.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=4|TP=3|ip-26-0-168-120]: Local number of parameters: 33.6M (64.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=6|TP=3|ip-26-0-168-238]: Local number of parameters: 67.1M (128.00MiB) +NCCL version 2.18.5+cuda12.2 +12/14/2024 23:38:36 [INFO|DP=0|PP=5|TP=3|ip-26-0-168-120]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=7|TP=3|ip-26-0-168-238]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5440.00MiB Peak reserved: 21826.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=2|TP=3|ip-26-0-162-14]: Local number of parameters: 33.6M (64.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=3|TP=3|ip-26-0-162-14]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=4|TP=3|ip-26-0-168-120]: [After model building] Memory usage: 64.03MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=6|TP=3|ip-26-0-168-238]: [After model building] Memory usage: 128.01MiB. Peak allocated: 5440.00MiB Peak reserved: 19778.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=2|TP=2|ip-26-0-162-14]: Local number of parameters: 33.6M (64.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=7|TP=2|ip-26-0-168-238]: Local number of parameters: 0 (0.00MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=2|TP=3|ip-26-0-162-14]: [After model building] Memory usage: 64.03MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=3|TP=2|ip-26-0-162-14]: Local number of parameters: 50.3M (96.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=6|TP=2|ip-26-0-168-238]: Local number of parameters: 67.1M (128.00MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=4|TP=2|ip-26-0-168-120]: Local number of parameters: 33.6M (64.02MiB) +NCCL version 2.18.5+cuda12.2 +12/14/2024 23:38:36 [INFO|DP=0|PP=2|TP=2|ip-26-0-162-14]: [After model building] Memory usage: 64.03MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=7|TP=2|ip-26-0-168-238]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5440.00MiB Peak reserved: 19778.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=7|TP=0|ip-26-0-168-238]: Local number of parameters: 0 (0.00MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=5|TP=2|ip-26-0-168-120]: Local number of parameters: 50.3M (96.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=2|TP=0|ip-26-0-162-14]: Local number of parameters: 33.6M (64.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=3|TP=0|ip-26-0-162-14]: Local number of parameters: 50.3M (96.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=3|TP=2|ip-26-0-162-14]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=6|TP=0|ip-26-0-168-238]: Local number of parameters: 67.1M (128.00MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=7|TP=1|ip-26-0-168-238]: Local number of parameters: 0 (0.00MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=5|TP=1|ip-26-0-168-120]: Local number of parameters: 50.3M (96.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=2|TP=1|ip-26-0-162-14]: Local number of parameters: 33.6M (64.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=4|TP=1|ip-26-0-168-120]: Local number of parameters: 33.6M (64.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=3|TP=1|ip-26-0-162-14]: Local number of parameters: 50.3M (96.02MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=6|TP=1|ip-26-0-168-238]: Local number of parameters: 67.1M (128.00MiB) +12/14/2024 23:38:36 [INFO|DP=0|PP=4|TP=2|ip-26-0-168-120]: [After model building] Memory usage: 64.03MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=6|TP=2|ip-26-0-168-238]: [After model building] Memory usage: 128.01MiB. Peak allocated: 5440.00MiB Peak reserved: 19778.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=7|TP=0|ip-26-0-168-238]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5440.00MiB Peak reserved: 19778.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=3|TP=0|ip-26-0-162-14]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=2|TP=0|ip-26-0-162-14]: [After model building] Memory usage: 64.03MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=5|TP=2|ip-26-0-168-120]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=6|TP=0|ip-26-0-168-238]: [After model building] Memory usage: 128.01MiB. Peak allocated: 5440.00MiB Peak reserved: 20802.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=5|TP=1|ip-26-0-168-120]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=7|TP=1|ip-26-0-168-238]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5440.00MiB Peak reserved: 19778.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=3|TP=1|ip-26-0-162-14]: [After model building] Memory usage: 96.04MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=6|TP=1|ip-26-0-168-238]: [After model building] Memory usage: 128.01MiB. Peak allocated: 5440.00MiB Peak reserved: 19778.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=4|TP=1|ip-26-0-168-120]: [After model building] Memory usage: 64.03MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/14/2024 23:38:36 [INFO|DP=0|PP=2|TP=1|ip-26-0-162-14]: [After model building] Memory usage: 64.03MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/14/2024 23:38:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: [Optimizer Building] Using LearningRateForSP as learning rate +12/14/2024 23:38:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/14/2024 23:38:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Using dummy data generator +12/14/2024 23:38:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: [Training Plan] There are 1 training stages +12/14/2024 23:38:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: [Stage Stable Training Stage] start from step 1 +12/14/2024 23:38:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: +12/14/2024 23:38:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: [Start training] datetime: 2024-12-14 23:38:38.797664 | mbs: 8 | grad_accum: 32 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/14/2024 23:38:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/14/2024 23:38:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-123]: Memory usage: 1120.13MiB. Peak allocated 5440.00MiB. Peak reserved: 22850.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +slurmstepd: error: *** JOB 13446329 ON ip-26-0-161-123 CANCELLED AT 2024-12-14T23:47:57 DUE TO TIME LIMIT *** +[2024-12-14 23:47:57,060] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 23:47:57,060] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108323 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100156 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108324 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100157 closing signal SIGTERM +slurmstepd: error: *** STEP 13446329.0 ON ip-26-0-161-123 CANCELLED AT 2024-12-14T23:47:57 DUE TO TIME LIMIT *** +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108325 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108326 closing signal SIGTERM +[2024-12-14 23:47:57,059] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 93254 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 93255 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 93256 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 93257 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 93258 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 93259 closing signal SIGTERM +[2024-12-14 23:47:57,061] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-14 23:47:57,061] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83064 closing signal SIGTERM +[2024-12-14 23:47:57,061] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83065 closing signal SIGTERM +[2024-12-14 23:47:57,061] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83066 closing signal SIGTERM +[2024-12-14 23:47:57,061] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83067 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100158 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100159 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100160 closing signal SIGTERM +[2024-12-14 23:47:57,061] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83068 closing signal SIGTERM +[2024-12-14 23:47:57,061] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83069 closing signal SIGTERM +[2024-12-14 23:47:57,061] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83070 closing signal SIGTERM +[2024-12-14 23:47:57,062] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83071 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108327 closing signal SIGTERM +[2024-12-14 23:47:57,062] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108328 closing signal SIGTERM +[2024-12-14 23:47:57,063] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108329 closing signal SIGTERM +[2024-12-14 23:47:57,063] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108330 closing signal SIGTERM +[2024-12-14 23:47:57,062] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100161 closing signal SIGTERM +[2024-12-14 23:47:57,060] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 93260 closing signal SIGTERM +[2024-12-14 23:47:57,061] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 93261 closing signal SIGTERM +[2024-12-14 23:47:57,065] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108323 closing signal SIGTERM +[2024-12-14 23:47:57,062] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100162 closing signal SIGTERM +[2024-12-14 23:47:57,065] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108324 closing signal SIGTERM +[2024-12-14 23:47:57,063] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100163 closing signal SIGTERM +[2024-12-14 23:47:57,065] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108325 closing signal SIGTERM +[2024-12-14 23:47:57,064] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100156 closing signal SIGTERM +[2024-12-14 23:47:57,065] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108326 closing signal SIGTERM +[2024-12-14 23:47:57,064] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100157 closing signal SIGTERM +[2024-12-14 23:47:57,065] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108327 closing signal SIGTERM +[2024-12-14 23:47:57,064] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100158 closing signal SIGTERM +[2024-12-14 23:47:57,065] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108328 closing signal SIGTERM +[2024-12-14 23:47:57,064] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100159 closing signal SIGTERM +[2024-12-14 23:47:57,065] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108329 closing signal SIGTERM +[2024-12-14 23:47:57,064] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100160 closing signal SIGTERM +[2024-12-14 23:47:57,065] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108330 closing signal SIGTERM +[2024-12-14 23:47:57,064] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100161 closing signal SIGTERM +[2024-12-14 23:47:57,064] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100162 closing signal SIGTERM +[2024-12-14 23:47:57,064] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100163 closing signal SIGTERM +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-14 23:47:57,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83064 closing signal SIGTERM +[2024-12-14 23:47:57,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83065 closing signal SIGTERM +[2024-12-14 23:47:57,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83066 closing signal SIGTERM +[2024-12-14 23:47:57,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83067 closing signal SIGTERM +[2024-12-14 23:47:57,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83068 closing signal SIGTERM +[2024-12-14 23:47:57,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83069 closing signal SIGTERM +[2024-12-14 23:47:57,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83070 closing signal SIGTERM +[2024-12-14 23:47:57,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83071 closing signal SIGTERM diff --git a/logs/13458789-bench_8.86G_dp8_tp1_pp4_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/13458789-bench_8.86G_dp8_tp1_pp4_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..de05e38c868525e5e1f2f6bed7d0a488674c4049 --- /dev/null +++ b/logs/13458789-bench_8.86G_dp8_tp1_pp4_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,2650 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n x ']' +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/slurm/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' +++ eval 'PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_SHLVL=1 ++++ CONDA_SHLVL=1 ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=2 +++ CONDA_SHLVL=2 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ scontrol show hostnames 'ip-26-0-162-180,ip-26-0-171-[21,56,88]' ++ export 'NODELIST=ip-26-0-162-180 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-88' ++ NODELIST='ip-26-0-162-180 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-88' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-162-180,ip-26-0-171-[21,56,88]' ++ export MASTER_NODE=ip-26-0-162-180 ++ MASTER_NODE=ip-26-0-162-180 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-162-180' +Master node: ip-26-0-162-180 ++ echo 'All nodes: ip-26-0-162-180 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-88' +All nodes: ip-26-0-162-180 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-88 ++ echo 'World size: 32' +World size: 32 ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13458789 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-162-180:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_8.86G_dp8_tp1_pp4_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-15 01:10:16,255] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 01:10:16,253] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 01:10:16,262] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 01:10:16,268] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 01:10:16,255] torch.distributed.run: [WARNING] +[2024-12-15 01:10:16,255] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 01:10:16,255] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 01:10:16,255] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 01:10:16,254] torch.distributed.run: [WARNING] +[2024-12-15 01:10:16,254] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 01:10:16,254] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 01:10:16,254] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 01:10:16,262] torch.distributed.run: [WARNING] +[2024-12-15 01:10:16,262] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 01:10:16,262] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 01:10:16,262] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 01:10:16,268] torch.distributed.run: [WARNING] +[2024-12-15 01:10:16,268] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 01:10:16,268] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 01:10:16,268] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +12/15/2024 01:10:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Measuring inter-GPU and intra-node bandwidth... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Bandwidth measurement complete. Time taken: 17.05 seconds +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Config: +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Config(general=GeneralArgs(project='debug', +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: run='8.86G_dp8_tp1_pp4_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k', +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: seed=42, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: step=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: consumed_train_samples=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: ignore_sanity_checks=True), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: parallelism=ParallelismArgs(dp=8, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pp=4, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tp=1, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pp_engine=, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tp_mode=, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tp_linear_async_communication=True, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: recompute_layer=False, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tp_recompute_allgather=True, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: expert_parallel_size=1), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: eos_token_id=0, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: hidden_act='silu', +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: hidden_size=4096, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: initializer_range=0.02, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: intermediate_size=14336, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: is_llama_config=True, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: max_position_embeddings=4096, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_attention_heads=32, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_hidden_layers=32, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_key_value_heads=32, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pad_token_id=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pretraining_tp=1, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rms_norm_eps=1e-05, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_scaling=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_theta=10000.0, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_interleaved=False, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tie_word_embeddings=False, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: use_cache=True, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: vocab_size=131072), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: init_method=RandomInit(std=0.02), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: dtype=torch.bfloat16, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: make_vocab_size_divisible_by=1, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: ddp_bucket_cap_mb=25), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tokenizer_revision=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tokenizer_max_length=None), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: checkpoint_interval=10000, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: save_initial_state=False, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: save_final_state=False, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: resume_checkpoint_path=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: checkpoints_path_is_shared_file_system=False), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: logging=LoggingArgs(log_level='info', +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: log_level_replica='info', +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: iteration_step_info_interval=1), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tokens=TokensArgs(sequence_length=4096, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: train_steps=100, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: micro_batch_size=8, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: batch_accumulation_per_replica=4, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: val_check_interval=100, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: limit_val_batches=0, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: limit_test_batches=0), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: adam_beta1=0.9, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: adam_beta2=0.95, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: torch_adam_is_fused=True, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: name='adamW'), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: zero_stage=0, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: weight_decay=0.01, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: clip_grad=1.0, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: accumulate_grad_in_fp32=True, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lr_warmup_steps=2, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lr_warmup_style='linear', +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lr_decay_style='cosine', +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lr_decay_steps=13, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lr_decay_starting_step=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: min_decay_lr=1e-05)), +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: start_training_step=1, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: data=DataArgs(dataset=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: seed=42, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_loading_workers=1))], +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: profiler=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lighteval=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: s3_upload=None) +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Model Config: +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: LlamaConfig(bos_token_id=0, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: eos_token_id=0, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: hidden_act='silu', +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: hidden_size=4096, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: initializer_range=0.02, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: intermediate_size=14336, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: is_llama_config=True, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: max_position_embeddings=4096, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_attention_heads=32, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_hidden_layers=32, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_key_value_heads=32, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pad_token_id=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pretraining_tp=1, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rms_norm_eps=1e-05, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_scaling=None, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_theta=10000.0, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_interleaved=False, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tie_word_embeddings=False, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: use_cache=True, +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: vocab_size=131072) +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Building model.. +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Initialize RoPE Theta = 10000.0 +12/15/2024 01:10:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Total number of parameters: 8.86G (16896.51MiB) +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Local number of parameters: 2.73G (5200.14MiB) +12/15/2024 01:11:00 [INFO|DP=0|PP=2|TP=0|ip-26-0-171-56]: Local number of parameters: 1.95G (3712.12MiB) +12/15/2024 01:11:00 [INFO|DP=0|PP=3|TP=0|ip-26-0-171-88]: Local number of parameters: 2G (3808.10MiB) +12/15/2024 01:11:00 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-21]: Local number of parameters: 2.19G (4176.14MiB) +12/15/2024 01:11:00 [INFO|DP=0|PP=2|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 3712.14MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/15/2024 01:11:00 [INFO|DP=0|PP=3|TP=0|ip-26-0-171-88]: [After model building] Memory usage: 3808.12MiB. Peak allocated: 5440.00MiB Peak reserved: 18754.00MiB +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: [After model building] Memory usage: 5200.16MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/15/2024 01:11:00 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-21]: [After model building] Memory usage: 4176.16MiB. Peak allocated: 5440.00MiB Peak reserved: 22850.00MiB +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: No checkpoint path provided. +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Parametrizing model parameters using StandardParametrizator +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: [Optimizer Building] Using LearningRateForSP as learning rate +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Using dummy data generator +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: [Training Plan] There are 1 training stages +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: [Stage Stable Training Stage] start from step 1 +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: [Start training] datetime: 2024-12-15 01:11:00.772795 | mbs: 8 | grad_accum: 4 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/15/2024 01:11:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Memory usage: 31200.86MiB. Peak allocated 31200.86MiB. Peak reserved: 48856.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + sharded_logits = self.model( File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 140, in forward +return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return F.linear(gathered_tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 2 has a total capacty of 79.33 GiB of which 827.94 MiB is free. Including non-PyTorch memory, this process has 78.51 GiB memory in use. Of the allocated memory 69.29 GiB is allocated by PyTorch, and 700.10 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 140, in forward + return F.linear(gathered_tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 0 has a total capacty of 79.33 GiB of which 859.94 MiB is free. Including non-PyTorch memory, this process has 78.48 GiB memory in use. Of the allocated memory 69.29 GiB is allocated by PyTorch, and 700.10 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 140, in forward + return F.linear(gathered_tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 3 has a total capacty of 79.33 GiB of which 707.94 MiB is free. Including non-PyTorch memory, this process has 78.63 GiB memory in use. Of the allocated memory 69.29 GiB is allocated by PyTorch, and 700.10 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward +Traceback (most recent call last): + output = model(**micro_batch)sharded_logits = self.model( + + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)return super().apply(*args, **kwargs) # type: ignore[misc] + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + trainer.train(dataloader)return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 140, in forward + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return F.linear(gathered_tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 5 has a total capacty of 79.33 GiB of which 827.94 MiB is free. Including non-PyTorch memory, this process has 78.51 GiB memory in use. Of the allocated memory 69.29 GiB is allocated by PyTorch, and 700.10 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self.module(*inputs, **kwargs) # type: ignore[index] +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) +return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 140, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return F.linear(gathered_tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 7 has a total capacty of 79.33 GiB of which 1.00 GiB is free. Including non-PyTorch memory, this process has 78.31 GiB memory in use. Of the allocated memory 69.29 GiB is allocated by PyTorch, and 700.10 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear +hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 140, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + return F.linear(gathered_tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 4 has a total capacty of 79.33 GiB of which 811.94 MiB is free. Including non-PyTorch memory, this process has 78.53 GiB memory in use. Of the allocated memory 69.29 GiB is allocated by PyTorch, and 700.10 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 140, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return F.linear(gathered_tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 1 has a total capacty of 79.33 GiB of which 827.94 MiB is free. Including non-PyTorch memory, this process has 78.51 GiB memory in use. Of the allocated memory 69.29 GiB is allocated by PyTorch, and 700.10 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 140, in forward + return F.linear(gathered_tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 6 has a total capacty of 79.33 GiB of which 827.94 MiB is free. Including non-PyTorch memory, this process has 78.51 GiB memory in use. Of the allocated memory 69.29 GiB is allocated by PyTorch, and 700.10 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train +trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train +trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train +trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + trainer.train(dataloader) + trainer.train(dataloader)trainer.train(dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + trainer.train(dataloader) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter +outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + outputs = self.pipeline_engine.train_batch_iter(outputs = self.pipeline_engine.train_batch_iter( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter(outputs = self.pipeline_engine.train_batch_iter( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + output = model(**micro_batch) +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward +return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + +return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return forward_call(*args, **kwargs) + +return forward_call(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + + + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + else self._run_ddp_forward(*inputs, **kwargs) + else self._run_ddp_forward(*inputs, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + else self._run_ddp_forward(*inputs, **kwargs)else self._run_ddp_forward(*inputs, **kwargs) + + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.module(*inputs, **kwargs) # type: ignore[index] return self.module(*inputs, **kwargs) # type: ignore[index] +return self.module(*inputs, **kwargs) # type: ignore[index] + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +else self._run_ddp_forward(*inputs, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +return self.module(*inputs, **kwargs) # type: ignore[index] File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + +return self.module(*inputs, **kwargs) # type: ignore[index] + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + trainer.train(dataloader) + trainer.train(dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train +trainer.train(dataloader) +trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train +trainer.train(dataloader) + trainer.train(dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train +outputs = self.pipeline_engine.train_batch_iter( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + else self._run_ddp_forward(*inputs, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + else self._run_ddp_forward(*inputs, **kwargs)else self._run_ddp_forward(*inputs, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward +else self._run_ddp_forward(*inputs, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + else self._run_ddp_forward(*inputs, **kwargs)return self.module(*inputs, **kwargs) # type: ignore[index] + + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + else self._run_ddp_forward(*inputs, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward +return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + return self.module(*inputs, **kwargs) # type: ignore[index] + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter(outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter(outputs = self.pipeline_engine.train_batch_iter(outputs = self.pipeline_engine.train_batch_iter( + + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 252, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 461, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 498, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + return self.module(*inputs, **kwargs) # type: ignore[index] + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 258, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + pg.recv([tensor], group_src_rank, tag).wait() +RuntimeError: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + output = model(**micro_batch) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + return self._call_impl(*args, **kwargs) +return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + return forward_call(*args, **kwargs) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward + else self._run_ddp_forward(*inputs, **kwargs) + else self._run_ddp_forward(*inputs, **kwargs) +else self._run_ddp_forward(*inputs, **kwargs) else self._run_ddp_forward(*inputs, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward +else self._run_ddp_forward(*inputs, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward +else self._run_ddp_forward(*inputs, **kwargs)else self._run_ddp_forward(*inputs, **kwargs) + + + + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + else self._run_ddp_forward(*inputs, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward +return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return self.module(*inputs, **kwargs) # type: ignore[index] + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return forward_call(*args, **kwargs) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward +return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.module(*inputs, **kwargs) # type: ignore[index] + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + sharded_logits = self.model( + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +sharded_logits = self.model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0]return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + sharded_logits = self.model( + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + return self.module(*inputs, **kwargs) # type: ignore[index] + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + hidden_encoder_states = encoder_block(**hidden_encoder_states) +return self.module(*inputs, **kwargs) # type: ignore[index] +return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +return self.module(*inputs, **kwargs) # type: ignore[index] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +return self.module(*inputs, **kwargs) # type: ignore[index] + return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + hidden_encoder_states = encoder_block(**hidden_encoder_states) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + hidden_encoder_states = encoder_block(**hidden_encoder_states) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + hidden_encoder_states = encoder_block(**hidden_encoder_states) + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.module(*inputs, **kwargs) # type: ignore[index] + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states)hidden_encoder_states = encoder_block(**hidden_encoder_states) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward +return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model(sharded_logits = self.model( + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +sharded_logits = self.model(return self._call_impl(*args, **kwargs) + + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + new_kwargs[name] = recv_from_pipeline_state_buffer( + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + new_kwargs[name] = recv_from_pipeline_state_buffer( File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + + new_kwargs[name] = recv_from_pipeline_state_buffer( File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + +new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + recv_activation_tensor = recv_activation() + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + new_kwargs[name] = recv_from_pipeline_state_buffer( File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + + new_kwargs[name] = recv_from_pipeline_state_buffer( File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer +new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer +new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication +pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +hidden_encoder_states = encoder_block(**hidden_encoder_states) + hidden_encoder_states = encoder_block(**hidden_encoder_states)hidden_encoder_states = encoder_block(**hidden_encoder_states) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication +pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication +pipeline_state.run_communication() +pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + pipeline_state.run_communication() + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + pipeline_state.run_communication() File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + recv_activation_tensor = recv_activation() +recv_activation_tensor = recv_activation() File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + + recv_activation_tensor = recv_activation() File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + recv_activation_tensor = recv_activation() + pipeline_state.run_communication()recv_activation_tensor = recv_activation() + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] +return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] +recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + meta = self._recv_meta(from_rank=from_rank, tag=tag) + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ +recv_activation_tensor = recv_activation() + recv_activation_tensor = recv_activation() + recv_activation_tensor = recv_activation() File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + return self._call_impl(*args, **kwargs) + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] +recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ +recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0]return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + +return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return self._call_impl(*args, **kwargs) + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward +return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + return forward_call(*args, **kwargs) + return func(*args, **kwargs) + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 126, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors +buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) +buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors +buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + meta = self._recv_meta(from_rank=from_rank, tag=tag) + meta = self._recv_meta(from_rank=from_rank, tag=tag) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta +meta = self._recv_meta(from_rank=from_rank, tag=tag) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta +meta = self._recv_meta(from_rank=from_rank, tag=tag) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.recv( +dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper +dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) +buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + pg.recv([tensor], group_src_rank, tag).wait() + meta = self._recv_meta(from_rank=from_rank, tag=tag) +meta = self._recv_meta(from_rank=from_rank, tag=tag) +RuntimeError File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta +: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + meta = self._recv_meta(from_rank=from_rank, tag=tag)dist.recv( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + new_kwargs[name] = recv_from_pipeline_state_buffer( + new_kwargs[name] = recv_from_pipeline_state_buffer( File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + + return func(*args, **kwargs) + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer +new_kwargs[name] = recv_from_pipeline_state_buffer( + new_kwargs[name] = recv_from_pipeline_state_buffer( File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 117, in recv_from_pipeline_state_buffer +pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + pipeline_state.run_communication() + pipeline_state.run_communication() File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + +pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication +pipeline_state.run_communication() File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + pipeline_state.run_communication() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 150, in run_communication + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv +return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + return func(*args, **kwargs) pg.recv([tensor], group_src_rank, tag).wait() +return func(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv +RuntimeError: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + return func(*args, **kwargs) + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv +return func(*args, **kwargs) + return func(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + return func(*args, **kwargs) + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv +return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + pg.recv([tensor], group_src_rank, tag).wait() +RuntimeError: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + pg.recv([tensor], group_src_rank, tag).wait() +RuntimeError: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ +recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ +recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ +recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ +recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ +recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ +recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ + pg.recv([tensor], group_src_rank, tag).wait() + recv_activation_tensor = recv_activation() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/state.py", line 31, in __call__ +RuntimeError: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + pg.recv([tensor], group_src_rank, tag).wait() +RuntimeError: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0]return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0]return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors +return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0]return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + return self.p2p.recv_tensors(num_tensors=1, from_rank=self.from_rank)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 353, in recv_tensors + pg.recv([tensor], group_src_rank, tag).wait() +RuntimeError: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + pg.recv([tensor], group_src_rank, tag).wait() + pg.recv([tensor], group_src_rank, tag).wait() +RuntimeError: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +RuntimeErrorpg.recv([tensor], group_src_rank, tag).wait(): + [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue.pg.recv([tensor], group_src_rank, tag).wait() + + RuntimeErrorpg.recv([tensor], group_src_rank, tag).wait()pg.recv([tensor], group_src_rank, tag).wait() : RuntimeError +[2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + +pg.recv([tensor], group_src_rank, tag).wait(): RuntimeError +[2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue.RuntimeError: : +[2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue.RuntimeError[2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +: +[2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + pg.recv([tensor], group_src_rank, tag).wait() +RuntimeError: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) +buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors +buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + +buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + +buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors +buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + buffers, futures = self.irecv_tensors(num_tensors=num_tensors, from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 326, in irecv_tensors + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + meta = self._recv_meta(from_rank=from_rank, tag=tag)meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + meta = self._recv_meta(from_rank=from_rank, tag=tag) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta +meta = self._recv_meta(from_rank=from_rank, tag=tag)meta = self._recv_meta(from_rank=from_rank, tag=tag) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/p2p.py", line 246, in _recv_meta + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper +dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.recv( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + return func(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv +return func(*args, **kwargs) + return func(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 1643, in recv + pg.recv([tensor], group_src_rank, tag).wait()pg.recv([tensor], group_src_rank, tag).wait()pg.recv([tensor], group_src_rank, tag).wait() + + +pg.recv([tensor], group_src_rank, tag).wait()pg.recv([tensor], group_src_rank, tag).wait() +pg.recv([tensor], group_src_rank, tag).wait() +pg.recv([tensor], group_src_rank, tag).wait() +RuntimeErrorRuntimeError +: RuntimeErrorRuntimeErrorRuntimeError: RuntimeError[1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue.RuntimeError: : : [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue.: +: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue.[1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue.[1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue.[1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +[1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. + + + + + pg.recv([tensor], group_src_rank, tag).wait() +RuntimeError: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +[2024-12-15 01:11:18,475] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61496 closing signal SIGTERM +[2024-12-15 01:11:18,475] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61497 closing signal SIGTERM +[2024-12-15 01:11:18,475] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61499 closing signal SIGTERM +[2024-12-15 01:11:18,475] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61500 closing signal SIGTERM +[2024-12-15 01:11:18,475] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61501 closing signal SIGTERM +[2024-12-15 01:11:18,475] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61502 closing signal SIGTERM +[2024-12-15 01:11:18,475] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61503 closing signal SIGTERM +[2024-12-15 01:11:22,208] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 2 (pid: 61498) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-15_01:11:18 + host : ip-26-0-162-180.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 61498) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-15 01:11:22,482] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-21.ec2.internal_83085_0' has failed to send a keep-alive heartbeat to the rendezvous '13458789' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-162-180: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13458789.0 +[2024-12-15 01:11:22,606] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 01:11:22,605] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 01:11:22,606] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 01:11:22,607] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 63303 closing signal SIGTERM +[2024-12-15 01:11:22,605] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83160 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39840 closing signal SIGTERM +[2024-12-15 01:11:22,607] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 63304 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83161 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39841 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83162 closing signal SIGTERM +[2024-12-15 01:11:22,607] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 63305 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39842 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83163 closing signal SIGTERM +[2024-12-15 01:11:22,607] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 63306 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39843 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83164 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83165 closing signal SIGTERM +[2024-12-15 01:11:22,607] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 63307 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39845 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39846 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83166 closing signal SIGTERM +[2024-12-15 01:11:22,606] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83167 closing signal SIGTERM +[2024-12-15 01:11:22,607] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 63308 closing signal SIGTERM +[2024-12-15 01:11:22,607] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 63309 closing signal SIGTERM +[2024-12-15 01:11:22,607] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 63310 closing signal SIGTERM +[2024-12-15 01:11:23,348] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-88.ec2.internal_39767_0' has failed to send a keep-alive heartbeat to the rendezvous '13458789' due to an error of type RendezvousConnectionError. +[2024-12-15 01:11:23,404] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-56.ec2.internal_63231_0' has failed to send a keep-alive heartbeat to the rendezvous '13458789' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 39767 got signal: 15 +srun: error: ip-26-0-171-88: task 3: Exited with exit code 1 +[2024-12-15 01:11:27,484] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-21.ec2.internal_83085_0' has failed to send a keep-alive heartbeat to the rendezvous '13458789' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 83085 got signal: 15 +srun: error: ip-26-0-171-21: task 1: Exited with exit code 1 +[2024-12-15 01:11:28,405] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-56.ec2.internal_63231_0' has failed to send a keep-alive heartbeat to the rendezvous '13458789' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 63231 got signal: 15 +srun: error: ip-26-0-171-56: task 2: Exited with exit code 1 diff --git a/logs/13458875-bench_1.34G_dp8_tp4_pp4_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/13458875-bench_1.34G_dp8_tp4_pp4_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..79be31cb652045a961e614709d5f350f8ad303da --- /dev/null +++ b/logs/13458875-bench_1.34G_dp8_tp4_pp4_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,1305 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n x ']' +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/slurm/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' +++ eval 'PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_SHLVL=1 ++++ CONDA_SHLVL=1 ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=2 +++ CONDA_SHLVL=2 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ scontrol show hostnames 'ip-26-0-160-[225,242],ip-26-0-161-[78,103,123],ip-26-0-165-[38,59,131,164],ip-26-0-168-[120,238],ip-26-0-169-[132,139],ip-26-0-173-[202,246],ip-26-0-174-36' ++ export 'NODELIST=ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-161-103 +ip-26-0-161-123 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36' ++ NODELIST='ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-161-103 +ip-26-0-161-123 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-160-[225,242],ip-26-0-161-[78,103,123],ip-26-0-165-[38,59,131,164],ip-26-0-168-[120,238],ip-26-0-169-[132,139],ip-26-0-173-[202,246],ip-26-0-174-36' ++ export MASTER_NODE=ip-26-0-160-225 ++ MASTER_NODE=ip-26-0-160-225 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=16 ++ NNODES=16 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=128 ++ WORLD_SIZE=128 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-160-225' +Master node: ip-26-0-160-225 ++ echo 'All nodes: ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-161-103 +ip-26-0-161-123 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36' +All nodes: ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-161-103 +ip-26-0-161-123 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36 ++ echo 'World size: 128' +World size: 128 ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=16 --nproc_per_node=8 --rdzv_id=13458875 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-225:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.34G_dp8_tp4_pp4_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,637] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,637] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,637] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,637] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,637] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,638] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,639] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,639] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,639] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,639] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,640] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,641] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,641] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,641] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,641] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,648] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,648] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,648] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,648] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,648] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,658] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,678] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,678] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,678] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,678] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,678] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,683] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,689] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,689] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,689] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,689] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,689] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,658] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,658] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,658] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,658] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,701] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,701] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,701] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,701] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,701] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,702] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,702] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,702] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,702] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,702] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,714] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,717] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,717] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,717] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,717] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,717] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,683] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,683] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,683] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,683] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,729] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 06:43:37,729] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,729] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,729] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,729] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,714] torch.distributed.run: [WARNING] +[2024-12-15 06:43:37,714] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 06:43:37,714] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 06:43:37,714] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +12/15/2024 06:44:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Measuring inter-GPU and intra-node bandwidth... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Bandwidth measurement complete. Time taken: 17.93 seconds +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Config: +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Config(general=GeneralArgs(project='debug', +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: run='1.34G_dp8_tp4_pp4_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k', +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: seed=42, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: step=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: consumed_train_samples=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: ignore_sanity_checks=True), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: parallelism=ParallelismArgs(dp=8, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pp=4, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp=4, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pp_engine=, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp_mode=, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp_linear_async_communication=True, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: recompute_layer=False, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tp_recompute_allgather=True, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: expert_parallel_size=1), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: eos_token_id=0, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_act='silu', +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_size=2048, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: initializer_range=0.02, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: intermediate_size=8192, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: is_llama_config=True, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: max_position_embeddings=4096, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_attention_heads=32, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_hidden_layers=16, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_key_value_heads=32, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pad_token_id=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pretraining_tp=1, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rms_norm_eps=1e-05, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_scaling=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_theta=10000.0, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_interleaved=False, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tie_word_embeddings=True, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: use_cache=True, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: vocab_size=131072), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: init_method=RandomInit(std=0.02), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: dtype=torch.bfloat16, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: make_vocab_size_divisible_by=1, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: ddp_bucket_cap_mb=25), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer_revision=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokenizer_max_length=None), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoint_interval=10000, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: save_initial_state=False, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: save_final_state=False, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: resume_checkpoint_path=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: checkpoints_path_is_shared_file_system=False), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: logging=LoggingArgs(log_level='info', +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: log_level_replica='info', +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: iteration_step_info_interval=1), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tokens=TokensArgs(sequence_length=4096, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: train_steps=100, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: micro_batch_size=8, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: batch_accumulation_per_replica=4, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: val_check_interval=100, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: limit_val_batches=0, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: limit_test_batches=0), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: adam_beta1=0.9, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: adam_beta2=0.95, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: torch_adam_is_fused=True, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: name='adamW'), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: zero_stage=0, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: weight_decay=0.01, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: clip_grad=1.0, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: accumulate_grad_in_fp32=True, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_warmup_steps=2, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_warmup_style='linear', +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_style='cosine', +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_steps=13, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lr_decay_starting_step=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: min_decay_lr=1e-05)), +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: start_training_step=1, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: data=DataArgs(dataset=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: seed=42, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_loading_workers=1))], +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: profiler=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: lighteval=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: s3_upload=None) +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Model Config: +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: LlamaConfig(bos_token_id=0, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: eos_token_id=0, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_act='silu', +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: hidden_size=2048, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: initializer_range=0.02, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: intermediate_size=8192, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: is_llama_config=True, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: max_position_embeddings=4096, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_attention_heads=32, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_hidden_layers=16, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: num_key_value_heads=32, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pad_token_id=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: pretraining_tp=1, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rms_norm_eps=1e-05, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_scaling=None, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_theta=10000.0, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: rope_interleaved=False, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: tie_word_embeddings=True, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: use_cache=True, +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: vocab_size=131072) +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Building model.. +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Initialize RoPE Theta = 10000.0 +12/15/2024 06:44:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/15/2024 06:44:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Total number of parameters: 1.61G (3072.52MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Local number of parameters: 168M (320.05MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [After model building] Memory usage: 320.06MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-225]: Local number of parameters: 168M (320.05MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=1|TP=0|ip-26-0-161-78]: Local number of parameters: 83.9M (160.04MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=3|TP=0|ip-26-0-169-139]: Local number of parameters: 67.1M (128.00MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-225]: Local number of parameters: 168M (320.05MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=2|TP=0|ip-26-0-165-59]: Local number of parameters: 83.9M (160.04MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-225]: [After model building] Memory usage: 320.06MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=3|TP=0|ip-26-0-169-139]: [After model building] Memory usage: 128.01MiB. Peak allocated: 5392.00MiB Peak reserved: 21782.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=1|TP=0|ip-26-0-161-78]: [After model building] Memory usage: 160.05MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-225]: [After model building] Memory usage: 320.06MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=2|TP=0|ip-26-0-165-59]: [After model building] Memory usage: 160.05MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-225]: Local number of parameters: 168M (320.05MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-225]: [After model building] Memory usage: 320.06MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=1|TP=1|ip-26-0-161-78]: Local number of parameters: 83.9M (160.04MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=2|TP=1|ip-26-0-165-59]: Local number of parameters: 83.9M (160.04MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=3|TP=1|ip-26-0-169-139]: Local number of parameters: 67.1M (128.00MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=1|TP=1|ip-26-0-161-78]: [After model building] Memory usage: 160.05MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=2|TP=1|ip-26-0-165-59]: [After model building] Memory usage: 160.05MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=2|TP=3|ip-26-0-165-59]: Local number of parameters: 83.9M (160.04MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=3|TP=3|ip-26-0-169-139]: Local number of parameters: 67.1M (128.00MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=3|TP=1|ip-26-0-169-139]: [After model building] Memory usage: 128.01MiB. Peak allocated: 5392.00MiB Peak reserved: 21782.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=1|TP=3|ip-26-0-161-78]: Local number of parameters: 83.9M (160.04MiB) +12/15/2024 06:44:27 [INFO|DP=0|PP=3|TP=3|ip-26-0-169-139]: [After model building] Memory usage: 128.01MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=2|TP=3|ip-26-0-165-59]: [After model building] Memory usage: 160.05MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=1|TP=3|ip-26-0-161-78]: [After model building] Memory usage: 160.05MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/15/2024 06:44:27 [INFO|DP=0|PP=1|TP=2|ip-26-0-161-78]: Local number of parameters: 83.9M (160.04MiB) +NCCL version 2.18.5+cuda12.2 +12/15/2024 06:44:27 [INFO|DP=0|PP=2|TP=2|ip-26-0-165-59]: Local number of parameters: 83.9M (160.04MiB) +NCCL version 2.18.5+cuda12.2 +12/15/2024 06:44:27 [INFO|DP=0|PP=1|TP=2|ip-26-0-161-78]: [After model building] Memory usage: 160.05MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +12/15/2024 06:44:27 [INFO|DP=0|PP=2|TP=2|ip-26-0-165-59]: [After model building] Memory usage: 160.05MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/15/2024 06:44:27 [INFO|DP=0|PP=3|TP=2|ip-26-0-169-139]: Local number of parameters: 67.1M (128.00MiB) +NCCL version 2.18.5+cuda12.2 +12/15/2024 06:44:27 [INFO|DP=0|PP=3|TP=2|ip-26-0-169-139]: [After model building] Memory usage: 128.01MiB. Peak allocated: 5392.00MiB Peak reserved: 22806.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/15/2024 06:44:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: No checkpoint path provided. +12/15/2024 06:44:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Parametrizing model parameters using StandardParametrizator +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/15/2024 06:44:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Optimizer Building] Using LearningRateForSP as learning rate +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/15/2024 06:44:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/15/2024 06:44:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Using dummy data generator +12/15/2024 06:44:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Training Plan] There are 1 training stages +12/15/2024 06:44:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Stage Stable Training Stage] start from step 1 +12/15/2024 06:44:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: +12/15/2024 06:44:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: [Start training] datetime: 2024-12-15 06:44:31.990020 | mbs: 8 | grad_accum: 4 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/15/2024 06:44:32 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/15/2024 06:44:32 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-225]: Memory usage: 1920.30MiB. Peak allocated 5392.00MiB. Peak reserved: 22806.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +slurmstepd: error: *** STEP 13458875.0 ON ip-26-0-160-225 CANCELLED AT 2024-12-15T06:53:32 DUE TO TIME LIMIT *** +[2024-12-15 06:53:32,380] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +slurmstepd: error: *** JOB 13458875 ON ip-26-0-160-225 CANCELLED AT 2024-12-15T06:53:32 DUE TO TIME LIMIT *** +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318118 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318119 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318120 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318121 closing signal SIGTERM +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-15 06:53:32,381] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165817 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165818 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165819 closing signal SIGTERM +[2024-12-15 06:53:32,380] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373741 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373742 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373743 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23565 closing signal SIGTERM +[2024-12-15 06:53:32,380] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184875 closing signal SIGTERM +[2024-12-15 06:53:32,380] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184876 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373744 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23566 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169624 closing signal SIGTERM +[2024-12-15 06:53:32,380] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184877 closing signal SIGTERM +[2024-12-15 06:53:32,380] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184878 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169625 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23567 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184879 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373745 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83812 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83813 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169626 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205458 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58228 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205459 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205460 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83814 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205461 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28884 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58229 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169628 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28885 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58230 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165820 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28886 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30875 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184880 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30876 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245313 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30877 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245314 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184881 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30878 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30879 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245315 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58231 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245316 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245317 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55634 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28887 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28888 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55635 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23568 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23569 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55636 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55637 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50826 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50827 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50828 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83815 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155642 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155643 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169629 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155644 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318122 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155645 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155646 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23570 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169630 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155647 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155648 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169631 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169632 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30880 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58232 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165821 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165822 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155649 closing signal SIGTERM +[2024-12-15 06:53:32,384] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83816 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98186 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245318 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245319 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245320 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98187 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50829 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50830 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50831 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50832 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98188 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98189 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98190 closing signal SIGTERM +[2024-12-15 06:53:32,385] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205462 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165823 closing signal SIGTERM +[2024-12-15 06:53:32,383] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165824 closing signal SIGTERM +[2024-12-15 06:53:32,384] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83817 closing signal SIGTERM +[2024-12-15 06:53:32,386] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83818 closing signal SIGTERM +[2024-12-15 06:53:32,385] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23571 closing signal SIGTERM +[2024-12-15 06:53:32,384] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28889 closing signal SIGTERM +[2024-12-15 06:53:32,385] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55638 closing signal SIGTERM +[2024-12-15 06:53:32,385] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23572 closing signal SIGTERM +[2024-12-15 06:53:32,386] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55639 closing signal SIGTERM +[2024-12-15 06:53:32,384] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28890 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23565 closing signal SIGTERM +[2024-12-15 06:53:32,386] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55640 closing signal SIGTERM +[2024-12-15 06:53:32,384] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28891 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23566 closing signal SIGTERM +[2024-12-15 06:53:32,386] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55641 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23567 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23568 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23569 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23570 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23571 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23572 closing signal SIGTERM +[2024-12-15 06:53:32,384] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30881 closing signal SIGTERM +[2024-12-15 06:53:32,384] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30882 closing signal SIGTERM +[2024-12-15 06:53:32,386] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98191 closing signal SIGTERM +[2024-12-15 06:53:32,385] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58233 closing signal SIGTERM +[2024-12-15 06:53:32,384] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318123 closing signal SIGTERM +[2024-12-15 06:53:32,385] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58234 closing signal SIGTERM +[2024-12-15 06:53:32,386] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58235 closing signal SIGTERM +[2024-12-15 06:53:32,386] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58228 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58229 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58230 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58231 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58232 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58233 closing signal SIGTERM +[2024-12-15 06:53:32,384] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318124 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58234 closing signal SIGTERM +[2024-12-15 06:53:32,384] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318125 closing signal SIGTERM +[2024-12-15 06:53:32,381] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184882 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58235 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318118 closing signal SIGTERM +[2024-12-15 06:53:32,382] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373746 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318119 closing signal SIGTERM +[2024-12-15 06:53:32,385] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205463 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318120 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318121 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318122 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318123 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318124 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 318125 closing signal SIGTERM +[2024-12-15 06:53:32,386] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205464 closing signal SIGTERM +[2024-12-15 06:53:32,385] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373747 closing signal SIGTERM +[2024-12-15 06:53:32,386] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205465 closing signal SIGTERM +[2024-12-15 06:53:32,385] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373748 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83812 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83813 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83814 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83815 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83816 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83817 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83818 closing signal SIGTERM +[2024-12-15 06:53:32,387] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83819 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98192 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55634 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28884 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55635 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55636 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28885 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55637 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28886 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55638 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28887 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55639 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28888 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55640 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28889 closing signal SIGTERM +[2024-12-15 06:53:32,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55641 closing signal SIGTERM +[2024-12-15 06:53:32,389] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98193 closing signal SIGTERM +[2024-12-15 06:53:32,389] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28890 closing signal SIGTERM +[2024-12-15 06:53:32,389] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28891 closing signal SIGTERM +[2024-12-15 06:53:32,389] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50833 closing signal SIGTERM +[2024-12-15 06:53:32,391] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50826 closing signal SIGTERM +[2024-12-15 06:53:32,391] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50827 closing signal SIGTERM +[2024-12-15 06:53:32,391] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50828 closing signal SIGTERM +[2024-12-15 06:53:32,391] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50829 closing signal SIGTERM +[2024-12-15 06:53:32,391] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50830 closing signal SIGTERM +[2024-12-15 06:53:32,391] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50831 closing signal SIGTERM +[2024-12-15 06:53:32,391] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50832 closing signal SIGTERM +[2024-12-15 06:53:32,391] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50833 closing signal SIGTERM +[2024-12-15 06:53:32,391] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98186 closing signal SIGTERM +[2024-12-15 06:53:32,391] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98187 closing signal SIGTERM +[2024-12-15 06:53:32,391] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98188 closing signal SIGTERM +[2024-12-15 06:53:32,392] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98189 closing signal SIGTERM +[2024-12-15 06:53:32,392] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98190 closing signal SIGTERM +[2024-12-15 06:53:32,392] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98191 closing signal SIGTERM +[2024-12-15 06:53:32,392] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98192 closing signal SIGTERM +[2024-12-15 06:53:32,392] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 98193 closing signal SIGTERM +[2024-12-15 06:53:32,412] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30875 closing signal SIGTERM +[2024-12-15 06:53:32,412] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30876 closing signal SIGTERM +[2024-12-15 06:53:32,412] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30877 closing signal SIGTERM +[2024-12-15 06:53:32,412] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30878 closing signal SIGTERM +[2024-12-15 06:53:32,412] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30879 closing signal SIGTERM +[2024-12-15 06:53:32,412] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30880 closing signal SIGTERM +[2024-12-15 06:53:32,412] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30881 closing signal SIGTERM +[2024-12-15 06:53:32,412] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30882 closing signal SIGTERM +[2024-12-15 06:53:32,432] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155642 closing signal SIGTERM +[2024-12-15 06:53:32,432] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155643 closing signal SIGTERM +[2024-12-15 06:53:32,432] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155644 closing signal SIGTERM +[2024-12-15 06:53:32,432] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155645 closing signal SIGTERM +[2024-12-15 06:53:32,432] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155646 closing signal SIGTERM +[2024-12-15 06:53:32,432] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155647 closing signal SIGTERM +[2024-12-15 06:53:32,432] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155648 closing signal SIGTERM +[2024-12-15 06:53:32,432] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155649 closing signal SIGTERM +[2024-12-15 06:53:32,508] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184875 closing signal SIGTERM +[2024-12-15 06:53:32,508] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184876 closing signal SIGTERM +[2024-12-15 06:53:32,508] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184877 closing signal SIGTERM +[2024-12-15 06:53:32,508] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184878 closing signal SIGTERM +[2024-12-15 06:53:32,508] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184879 closing signal SIGTERM +[2024-12-15 06:53:32,508] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184880 closing signal SIGTERM +[2024-12-15 06:53:32,508] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184881 closing signal SIGTERM +[2024-12-15 06:53:32,508] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184882 closing signal SIGTERM +[2024-12-15 06:53:32,524] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169624 closing signal SIGTERM +[2024-12-15 06:53:32,525] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169625 closing signal SIGTERM +[2024-12-15 06:53:32,525] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169626 closing signal SIGTERM +[2024-12-15 06:53:32,525] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169628 closing signal SIGTERM +[2024-12-15 06:53:32,525] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169629 closing signal SIGTERM +[2024-12-15 06:53:32,525] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169630 closing signal SIGTERM +[2024-12-15 06:53:32,525] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169631 closing signal SIGTERM +[2024-12-15 06:53:32,525] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169632 closing signal SIGTERM +[2024-12-15 06:53:32,553] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245313 closing signal SIGTERM +[2024-12-15 06:53:32,553] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245314 closing signal SIGTERM +[2024-12-15 06:53:32,553] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245315 closing signal SIGTERM +[2024-12-15 06:53:32,553] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245316 closing signal SIGTERM +[2024-12-15 06:53:32,553] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245317 closing signal SIGTERM +[2024-12-15 06:53:32,553] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245318 closing signal SIGTERM +[2024-12-15 06:53:32,553] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245319 closing signal SIGTERM +[2024-12-15 06:53:32,553] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 245320 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205458 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205459 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205460 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205461 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205462 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205463 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205464 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373741 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205465 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373742 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373743 closing signal SIGTERM +[2024-12-15 06:53:32,604] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373744 closing signal SIGTERM +[2024-12-15 06:53:32,605] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373745 closing signal SIGTERM +[2024-12-15 06:53:32,605] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373746 closing signal SIGTERM +[2024-12-15 06:53:32,605] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373747 closing signal SIGTERM +[2024-12-15 06:53:32,605] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 373748 closing signal SIGTERM +[2024-12-15 06:53:32,611] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165817 closing signal SIGTERM +[2024-12-15 06:53:32,611] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165818 closing signal SIGTERM +[2024-12-15 06:53:32,611] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165819 closing signal SIGTERM +[2024-12-15 06:53:32,611] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165820 closing signal SIGTERM +[2024-12-15 06:53:32,611] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165821 closing signal SIGTERM +[2024-12-15 06:53:32,611] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165822 closing signal SIGTERM +[2024-12-15 06:53:32,612] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165823 closing signal SIGTERM +[2024-12-15 06:53:32,612] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 165824 closing signal SIGTERM diff --git a/logs/13459859-bench_469G_dp1_tp1_pp16_acc256_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/13459859-bench_469G_dp1_tp1_pp16_acc256_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..ca1d022fe7d93dbf2b01d68c9bac2415c4125cea --- /dev/null +++ b/logs/13459859-bench_469G_dp1_tp1_pp16_acc256_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,665 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n x ']' +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/slurm/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' +++ eval 'PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_SHLVL=1 ++++ CONDA_SHLVL=1 ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=2 +++ CONDA_SHLVL=2 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ scontrol show hostnames 'ip-26-0-169-[132,139]' ++ export 'NODELIST=ip-26-0-169-132 +ip-26-0-169-139' ++ NODELIST='ip-26-0-169-132 +ip-26-0-169-139' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-169-[132,139]' ++ export MASTER_NODE=ip-26-0-169-132 ++ MASTER_NODE=ip-26-0-169-132 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-169-132' +Master node: ip-26-0-169-132 ++ echo 'All nodes: ip-26-0-169-132 +ip-26-0-169-139' +All nodes: ip-26-0-169-132 +ip-26-0-169-139 ++ echo 'World size: 16' +World size: 16 ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13459859 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-169-132:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_469G_dp1_tp1_pp16_acc256_mbs1_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-15 12:22:40,911] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 12:22:40,911] torch.distributed.run: [WARNING] +[2024-12-15 12:22:40,911] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 12:22:40,911] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 12:22:40,911] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 12:22:41,075] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 12:22:41,075] torch.distributed.run: [WARNING] +[2024-12-15 12:22:41,075] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 12:22:41,075] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 12:22:41,075] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +12/15/2024 12:23:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Measuring inter-GPU and intra-node bandwidth... +NCCL version 2.18.5+cuda12.2 +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Bandwidth measurement complete. Time taken: 16.31 seconds +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Config: +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Config(general=GeneralArgs(project='debug', +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: run='469G_dp1_tp1_pp16_acc256_mbs1_seq4096_zero0_tpmodeRED_vocab131k', +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: seed=42, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: step=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: consumed_train_samples=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: ignore_sanity_checks=True), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: parallelism=ParallelismArgs(dp=1, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pp=16, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp=1, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pp_engine=, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp_mode=, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp_linear_async_communication=True, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: recompute_layer=False, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tp_recompute_allgather=True, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: expert_parallel_size=1), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: eos_token_id=0, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_act='silu', +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_size=16384, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: initializer_range=0.02, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: intermediate_size=53248, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: is_llama_config=True, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: max_position_embeddings=4096, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_attention_heads=128, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_hidden_layers=126, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_key_value_heads=128, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pad_token_id=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pretraining_tp=1, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rms_norm_eps=1e-05, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_scaling=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_theta=10000.0, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_interleaved=False, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tie_word_embeddings=False, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: use_cache=True, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: vocab_size=131072), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: init_method=RandomInit(std=0.02), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: dtype=torch.bfloat16, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: make_vocab_size_divisible_by=1, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: ddp_bucket_cap_mb=25), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokenizer_revision=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokenizer_max_length=None), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: checkpoint_interval=10000, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: save_initial_state=False, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: save_final_state=False, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: resume_checkpoint_path=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: checkpoints_path_is_shared_file_system=False), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: logging=LoggingArgs(log_level='info', +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: log_level_replica='info', +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: iteration_step_info_interval=1), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tokens=TokensArgs(sequence_length=4096, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: train_steps=100, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: micro_batch_size=1, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: batch_accumulation_per_replica=256, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: val_check_interval=100, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: limit_val_batches=0, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: limit_test_batches=0), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: adam_beta1=0.9, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: adam_beta2=0.95, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: torch_adam_is_fused=True, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: name='adamW'), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: zero_stage=0, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: weight_decay=0.01, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: clip_grad=1.0, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: accumulate_grad_in_fp32=True, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_warmup_steps=2, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_warmup_style='linear', +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_decay_style='cosine', +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_decay_steps=13, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lr_decay_starting_step=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: min_decay_lr=1e-05)), +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: start_training_step=1, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: data=DataArgs(dataset=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: seed=42, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_loading_workers=1))], +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: profiler=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: lighteval=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: s3_upload=None) +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Model Config: +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: LlamaConfig(bos_token_id=0, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: eos_token_id=0, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_act='silu', +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: hidden_size=16384, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: initializer_range=0.02, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: intermediate_size=53248, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: is_llama_config=True, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: max_position_embeddings=4096, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_attention_heads=128, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_hidden_layers=126, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: num_key_value_heads=128, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pad_token_id=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: pretraining_tp=1, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rms_norm_eps=1e-05, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_scaling=None, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_theta=10000.0, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: rope_interleaved=False, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: tie_word_embeddings=False, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: use_cache=True, +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: vocab_size=131072) +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Building model.. +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Initialize RoPE Theta = 10000.0 +12/15/2024 12:23:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-169-132]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 + +ip-26-0-169-139:189395:189694 [6] transport/nvls.cc:155 NCCL WARN Cuda failure 'out of memory' + +ip-26-0-169-139:189394:189697 [5] transport/nvls.cc:155 NCCL WARN Cuda failure 'out of memory' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 248, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 170, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 170, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 718, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 718, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 728, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 824, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 728, in _init_model_instance + dist.all_reduce(total_params, group=parallel_context.pp_pg, async_op=False, op=dist.ReduceOp.SUM) # PP + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 824, in _init_model + dist.all_reduce(total_params, group=parallel_context.pp_pg, async_op=False, op=dist.ReduceOp.SUM) # PP + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + work = group.allreduce([tensor], opts) +torch.distributed.DistBackendError: NCCL error in: /opt/conda/conda-bld/pytorch_1699449201336/work/torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1333, unhandled cuda error (run with NCCL_DEBUG=INFO for details), NCCL version 2.18.5 +ncclUnhandledCudaError: Call to CUDA function failed. +Last error: +Cuda failure 'out of memory' + work = group.allreduce([tensor], opts) +torch.distributed.DistBackendError: NCCL error in: /opt/conda/conda-bld/pytorch_1699449201336/work/torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1333, unhandled cuda error (run with NCCL_DEBUG=INFO for details), NCCL version 2.18.5 +ncclUnhandledCudaError: Call to CUDA function failed. +Last error: +Cuda failure 'out of memory' +[2024-12-15 12:23:42,117] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189389 closing signal SIGTERM +[2024-12-15 12:23:42,117] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189390 closing signal SIGTERM +[2024-12-15 12:23:42,117] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189391 closing signal SIGTERM +[2024-12-15 12:23:42,117] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189392 closing signal SIGTERM +[2024-12-15 12:23:42,117] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189393 closing signal SIGTERM +[2024-12-15 12:23:42,117] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189396 closing signal SIGTERM +[2024-12-15 12:23:57,737] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 5 (pid: 189394) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-15_12:23:42 + host : ip-26-0-169-139.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 189395) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-15_12:23:42 + host : ip-26-0-169-139.ec2.internal + rank : 13 (local_rank: 5) + exitcode : 1 (pid: 189394) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-169-139: task 1: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13459859.0 +slurmstepd: error: *** STEP 13459859.0 ON ip-26-0-169-132 CANCELLED AT 2024-12-15T12:23:58 *** +[2024-12-15 12:23:58,032] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 12:23:58,033] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 117843 closing signal SIGTERM +[2024-12-15 12:23:58,033] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 117844 closing signal SIGTERM +[2024-12-15 12:23:58,033] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 117845 closing signal SIGTERM +[2024-12-15 12:23:58,033] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 117846 closing signal SIGTERM +[2024-12-15 12:23:58,033] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 117847 closing signal SIGTERM +[2024-12-15 12:23:58,033] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 117848 closing signal SIGTERM +[2024-12-15 12:23:58,034] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 117849 closing signal SIGTERM +[2024-12-15 12:23:58,034] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 117850 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 33, in + sys.exit(load_entry_point('torch==2.1.1', 'console_scripts', 'torchrun')()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 117771 got signal: 15 +srun: error: ip-26-0-169-132: task 0: Exited with exit code 1 diff --git a/logs/13459879-bench_8.86G_dp2_tp1_pp8_acc128_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/13459879-bench_8.86G_dp2_tp1_pp8_acc128_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..71bbcdf04987c10a4d48af4f0a22ab4ef9e596b4 --- /dev/null +++ b/logs/13459879-bench_8.86G_dp2_tp1_pp8_acc128_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,641 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n x ']' +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/slurm/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' +++ eval 'PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_SHLVL=1 ++++ CONDA_SHLVL=1 ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=2 +++ CONDA_SHLVL=2 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ scontrol show hostnames 'ip-26-0-171-[21,56]' ++ export 'NODELIST=ip-26-0-171-21 +ip-26-0-171-56' ++ NODELIST='ip-26-0-171-21 +ip-26-0-171-56' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-171-[21,56]' ++ export MASTER_NODE=ip-26-0-171-21 ++ MASTER_NODE=ip-26-0-171-21 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-43 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 52-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 52-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 52-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 52-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ echo 'Master node: ip-26-0-171-21' +Master node: ip-26-0-171-21 ++ echo 'All nodes: ip-26-0-171-21 +ip-26-0-171-56' +All nodes: ip-26-0-171-21 +ip-26-0-171-56 ++ echo 'World size: 16' +World size: 16 ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13459879 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-171-21:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_8.86G_dp2_tp1_pp8_acc128_mbs1_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-15 12:33:25,818] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 12:33:25,818] torch.distributed.run: [WARNING] +[2024-12-15 12:33:25,818] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 12:33:25,818] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 12:33:25,818] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 12:33:26,382] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-15 12:33:26,382] torch.distributed.run: [WARNING] +[2024-12-15 12:33:26,382] torch.distributed.run: [WARNING] ***************************************** +[2024-12-15 12:33:26,382] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-15 12:33:26,382] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +12/15/2024 12:33:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Measuring inter-GPU and intra-node bandwidth... +NCCL version 2.18.5+cuda12.2 +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Bandwidth measurement complete. Time taken: 15.34 seconds +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Config: +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Config(general=GeneralArgs(project='debug', +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: run='8.86G_dp2_tp1_pp8_acc128_mbs1_seq4096_zero0_tpmodeRED_vocab131k', +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: seed=42, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: step=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: consumed_train_samples=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: ignore_sanity_checks=True), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: parallelism=ParallelismArgs(dp=2, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: pp=8, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: tp=1, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: pp_engine=, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: tp_mode=, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: tp_linear_async_communication=True, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: recompute_layer=False, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: tp_recompute_allgather=True, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: expert_parallel_size=1), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: eos_token_id=0, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: hidden_act='silu', +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: hidden_size=4096, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: initializer_range=0.02, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: intermediate_size=14336, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: is_llama_config=True, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: max_position_embeddings=4096, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: num_attention_heads=32, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: num_hidden_layers=32, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: num_key_value_heads=32, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: pad_token_id=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: pretraining_tp=1, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: rms_norm_eps=1e-05, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: rope_scaling=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: rope_theta=10000.0, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: rope_interleaved=False, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: tie_word_embeddings=False, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: use_cache=True, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: vocab_size=131072), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: init_method=RandomInit(std=0.02), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: dtype=torch.bfloat16, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: make_vocab_size_divisible_by=1, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: ddp_bucket_cap_mb=25), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: tokenizer_revision=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: tokenizer_max_length=None), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: checkpoint_interval=10000, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: save_initial_state=False, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: save_final_state=False, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: resume_checkpoint_path=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: checkpoints_path_is_shared_file_system=False), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: logging=LoggingArgs(log_level='info', +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: log_level_replica='info', +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: iteration_step_info_interval=1), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: tokens=TokensArgs(sequence_length=4096, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: train_steps=100, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: micro_batch_size=1, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: batch_accumulation_per_replica=128, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: val_check_interval=100, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: limit_val_batches=0, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: limit_test_batches=0), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: adam_beta1=0.9, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: adam_beta2=0.95, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: torch_adam_is_fused=True, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: name='adamW'), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: zero_stage=0, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: weight_decay=0.01, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: clip_grad=1.0, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: accumulate_grad_in_fp32=True, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: lr_warmup_steps=2, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: lr_warmup_style='linear', +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: lr_decay_style='cosine', +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: lr_decay_steps=13, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: lr_decay_starting_step=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: min_decay_lr=1e-05)), +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: start_training_step=1, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: data=DataArgs(dataset=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: seed=42, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: num_loading_workers=1))], +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: profiler=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: lighteval=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: s3_upload=None) +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Model Config: +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: LlamaConfig(bos_token_id=0, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: eos_token_id=0, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: hidden_act='silu', +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: hidden_size=4096, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: initializer_range=0.02, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: intermediate_size=14336, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: is_llama_config=True, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: max_position_embeddings=4096, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: num_attention_heads=32, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: num_hidden_layers=32, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: num_key_value_heads=32, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: pad_token_id=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: pretraining_tp=1, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: rms_norm_eps=1e-05, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: rope_scaling=None, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: rope_theta=10000.0, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: rope_interleaved=False, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: tie_word_embeddings=False, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: use_cache=True, +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: vocab_size=131072) +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Building model.. +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Initialize RoPE Theta = 10000.0 +12/15/2024 12:34:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/15/2024 12:34:14 [INFO|DP=0|PP=4|TP=0|ip-26-0-171-56]: Local number of parameters: 973M (1856.06MiB) +12/15/2024 12:34:14 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-21]: Local number of parameters: 973M (1856.06MiB) +12/15/2024 12:34:14 [INFO|DP=0|PP=6|TP=0|ip-26-0-171-56]: Local number of parameters: 973M (1856.06MiB) +12/15/2024 12:34:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Total number of parameters: 8.86G (16896.51MiB) +12/15/2024 12:34:14 [INFO|DP=0|PP=3|TP=0|ip-26-0-171-21]: Local number of parameters: 1.22G (2320.08MiB) +12/15/2024 12:34:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Local number of parameters: 1.75G (3344.08MiB) +12/15/2024 12:34:14 [INFO|DP=0|PP=5|TP=0|ip-26-0-171-56]: Local number of parameters: 973M (1856.06MiB) +12/15/2024 12:34:14 [INFO|DP=0|PP=2|TP=0|ip-26-0-171-21]: Local number of parameters: 973M (1856.06MiB) +12/15/2024 12:34:14 [INFO|DP=0|PP=7|TP=0|ip-26-0-171-56]: Local number of parameters: 1.02G (1952.04MiB) +12/15/2024 12:34:14 [INFO|DP=0|PP=4|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 1856.08MiB. Peak allocated: 5504.00MiB Peak reserved: 22914.00MiB +12/15/2024 12:34:14 [INFO|DP=0|PP=3|TP=0|ip-26-0-171-21]: [After model building] Memory usage: 2320.09MiB. Peak allocated: 5504.00MiB Peak reserved: 16770.00MiB +12/15/2024 12:34:14 [INFO|DP=0|PP=6|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 1856.08MiB. Peak allocated: 5504.00MiB Peak reserved: 22914.00MiB +12/15/2024 12:34:14 [INFO|DP=0|PP=7|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 1952.05MiB. Peak allocated: 5504.00MiB Peak reserved: 22914.00MiB +12/15/2024 12:34:14 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-21]: [After model building] Memory usage: 1856.08MiB. Peak allocated: 5504.00MiB Peak reserved: 19842.00MiB +12/15/2024 12:34:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: [After model building] Memory usage: 3344.09MiB. Peak allocated: 5504.00MiB Peak reserved: 16770.00MiB +12/15/2024 12:34:14 [INFO|DP=0|PP=5|TP=0|ip-26-0-171-56]: [After model building] Memory usage: 1856.08MiB. Peak allocated: 5504.00MiB Peak reserved: 22914.00MiB +12/15/2024 12:34:14 [INFO|DP=0|PP=2|TP=0|ip-26-0-171-21]: [After model building] Memory usage: 1856.08MiB. Peak allocated: 5504.00MiB Peak reserved: 17794.00MiB +12/15/2024 12:34:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: No checkpoint path provided. +12/15/2024 12:34:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Parametrizing model parameters using StandardParametrizator +12/15/2024 12:34:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: [Optimizer Building] Using LearningRateForSP as learning rate +12/15/2024 12:34:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/15/2024 12:34:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Using dummy data generator +12/15/2024 12:34:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: [Training Plan] There are 1 training stages +12/15/2024 12:34:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: [Stage Stable Training Stage] start from step 1 +12/15/2024 12:34:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: +12/15/2024 12:34:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: [Start training] datetime: 2024-12-15 12:34:17.145148 | mbs: 1 | grad_accum: 128 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +12/15/2024 12:34:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/15/2024 12:34:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-21]: Memory usage: 20064.48MiB. Peak allocated 20064.48MiB. Peak reserved: 33496.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +slurmstepd: error: *** JOB 13459879 ON ip-26-0-171-21 CANCELLED AT 2024-12-15T12:43:16 DUE TO TIME LIMIT *** +[2024-12-15 12:43:16,801] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 12:43:16,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188644 closing signal SIGTERM +[2024-12-15 12:43:16,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188645 closing signal SIGTERM +[2024-12-15 12:43:16,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188646 closing signal SIGTERM +[2024-12-15 12:43:16,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188647 closing signal SIGTERM +slurmstepd: error: *** STEP 13459879.0 ON ip-26-0-171-21 CANCELLED AT 2024-12-15T12:43:16 DUE TO TIME LIMIT *** +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-15 12:43:16,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188648 closing signal SIGTERM +[2024-12-15 12:43:16,803] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-15 12:43:16,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72229 closing signal SIGTERM +[2024-12-15 12:43:16,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72230 closing signal SIGTERM +[2024-12-15 12:43:16,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72231 closing signal SIGTERM +[2024-12-15 12:43:16,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72232 closing signal SIGTERM +[2024-12-15 12:43:16,804] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72233 closing signal SIGTERM +[2024-12-15 12:43:16,804] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72234 closing signal SIGTERM +[2024-12-15 12:43:16,804] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188649 closing signal SIGTERM +[2024-12-15 12:43:16,805] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72235 closing signal SIGTERM +[2024-12-15 12:43:16,805] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188650 closing signal SIGTERM +[2024-12-15 12:43:16,805] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188651 closing signal SIGTERM +[2024-12-15 12:43:16,808] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72236 closing signal SIGTERM +[2024-12-15 12:43:16,808] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72229 closing signal SIGTERM +[2024-12-15 12:43:16,808] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72230 closing signal SIGTERM +[2024-12-15 12:43:16,808] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72231 closing signal SIGTERM +[2024-12-15 12:43:16,808] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72232 closing signal SIGTERM +[2024-12-15 12:43:16,808] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72233 closing signal SIGTERM +[2024-12-15 12:43:16,808] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72234 closing signal SIGTERM +[2024-12-15 12:43:16,808] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72235 closing signal SIGTERM +[2024-12-15 12:43:16,808] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72236 closing signal SIGTERM +[2024-12-15 12:43:16,860] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188644 closing signal SIGTERM +[2024-12-15 12:43:16,861] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188645 closing signal SIGTERM +[2024-12-15 12:43:16,861] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188646 closing signal SIGTERM +[2024-12-15 12:43:16,861] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188647 closing signal SIGTERM +[2024-12-15 12:43:16,861] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188648 closing signal SIGTERM +[2024-12-15 12:43:16,861] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188649 closing signal SIGTERM +[2024-12-15 12:43:16,861] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188650 closing signal SIGTERM +[2024-12-15 12:43:16,861] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 188651 closing signal SIGTERM diff --git a/logs/13502136-bench_1.34G_dp8_tp16_pp1_acc2_mbs16_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/13502136-bench_1.34G_dp8_tp16_pp1_acc2_mbs16_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..e6aa49d2cb5c978efb53a1701bc21f93666e1fc2 --- /dev/null +++ b/logs/13502136-bench_1.34G_dp8_tp16_pp1_acc2_mbs16_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,1466 @@ ++ '[' -z 13502136 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/slurm/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/opt/slurm/bin:1:/fsx/nouamane/miniconda/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/opt/slurm/bin:1:/fsx/nouamane/miniconda/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-163-[147,226],ip-26-0-164-[18,236],ip-26-0-165-[24,38],ip-26-0-168-52,ip-26-0-171-88,ip-26-0-172-142,ip-26-0-174-[100,186,196,240],ip-26-0-175-[34,132,165]' ++ export 'NODELIST=ip-26-0-163-147 +ip-26-0-163-226 +ip-26-0-164-18 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-168-52 +ip-26-0-171-88 +ip-26-0-172-142 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165' ++ NODELIST='ip-26-0-163-147 +ip-26-0-163-226 +ip-26-0-164-18 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-168-52 +ip-26-0-171-88 +ip-26-0-172-142 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-163-[147,226],ip-26-0-164-[18,236],ip-26-0-165-[24,38],ip-26-0-168-52,ip-26-0-171-88,ip-26-0-172-142,ip-26-0-174-[100,186,196,240],ip-26-0-175-[34,132,165]' ++ export MASTER_NODE=ip-26-0-163-147 ++ MASTER_NODE=ip-26-0-163-147 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=16 ++ NNODES=16 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=128 ++ WORLD_SIZE=128 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-163-147' +Master node: ip-26-0-163-147 ++ echo 'All nodes: ip-26-0-163-147 +ip-26-0-163-226 +ip-26-0-164-18 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-168-52 +ip-26-0-171-88 +ip-26-0-172-142 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165' +All nodes: ip-26-0-163-147 +ip-26-0-163-226 +ip-26-0-164-18 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-168-52 +ip-26-0-171-88 +ip-26-0-172-142 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165 ++ echo 'World size: 128' +World size: 128 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=16 --nproc_per_node=8 --rdzv_id=13502136 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-163-147:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.34G_dp8_tp16_pp1_acc2_mbs16_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,251] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,249] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,254] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,257] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,257] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,261] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,249] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,249] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,249] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,249] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,253] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,257] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,257] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,257] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,257] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,257] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,257] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,257] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,257] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,254] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,254] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,254] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,254] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,261] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,261] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,261] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,261] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,307] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 02:04:04,307] torch.distributed.run: [WARNING] +[2024-12-19 02:04:04,307] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 02:04:04,307] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 02:04:04,307] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +12/19/2024 02:04:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Measuring inter-GPU and intra-node bandwidth... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Bandwidth measurement complete. Time taken: 17.92 seconds +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Config: +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Config(general=GeneralArgs(project='debug', +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: run='1.34G_dp8_tp16_pp1_acc2_mbs16_seq4096_zero0_tpmodeRED_vocab131k', +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: seed=42, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: step=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: consumed_train_samples=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: ignore_sanity_checks=True), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: parallelism=ParallelismArgs(dp=8, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: pp=1, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: tp=16, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: pp_engine=, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: tp_mode=, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: tp_linear_async_communication=True, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: recompute_layer=False, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: tp_recompute_allgather=True, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: expert_parallel_size=1), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: eos_token_id=0, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: hidden_act='silu', +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: hidden_size=2048, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: initializer_range=0.02, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: intermediate_size=8192, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: is_llama_config=True, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: max_position_embeddings=4096, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: num_attention_heads=32, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: num_hidden_layers=16, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: num_key_value_heads=32, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: pad_token_id=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: pretraining_tp=1, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: rms_norm_eps=1e-05, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: rope_scaling=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: rope_theta=10000.0, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: rope_interleaved=False, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: tie_word_embeddings=True, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: use_cache=True, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: vocab_size=131072), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: init_method=RandomInit(std=0.02), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: dtype=torch.bfloat16, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: make_vocab_size_divisible_by=1, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: ddp_bucket_cap_mb=25), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: tokenizer_revision=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: tokenizer_max_length=None), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: checkpoint_interval=10000, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: save_initial_state=False, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: save_final_state=False, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: resume_checkpoint_path=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: checkpoints_path_is_shared_file_system=False), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: logging=LoggingArgs(log_level='info', +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: log_level_replica='info', +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: iteration_step_info_interval=1), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: tokens=TokensArgs(sequence_length=4096, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: train_steps=100, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: micro_batch_size=16, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: batch_accumulation_per_replica=2, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: val_check_interval=100, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: limit_val_batches=0, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: limit_test_batches=0), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: adam_beta1=0.9, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: adam_beta2=0.95, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: torch_adam_is_fused=True, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: name='adamW'), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: zero_stage=0, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: weight_decay=0.01, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: clip_grad=1.0, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: accumulate_grad_in_fp32=True, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: lr_warmup_steps=2, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: lr_warmup_style='linear', +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: lr_decay_style='cosine', +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: lr_decay_steps=13, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: lr_decay_starting_step=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: min_decay_lr=1e-05)), +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: start_training_step=1, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: data=DataArgs(dataset=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: seed=42, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: num_loading_workers=1))], +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: profiler=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: lighteval=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: s3_upload=None) +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Model Config: +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: LlamaConfig(bos_token_id=0, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: eos_token_id=0, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: hidden_act='silu', +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: hidden_size=2048, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: initializer_range=0.02, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: intermediate_size=8192, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: is_llama_config=True, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: max_position_embeddings=4096, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: num_attention_heads=32, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: num_hidden_layers=16, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: num_key_value_heads=32, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: pad_token_id=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: pretraining_tp=1, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: rms_norm_eps=1e-05, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: rope_scaling=None, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: rope_theta=10000.0, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: rope_interleaved=False, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: tie_word_embeddings=True, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: use_cache=True, +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: vocab_size=131072) +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Building model.. +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Initialize RoPE Theta = 10000.0 +12/19/2024 02:04:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Total number of parameters: 1.34G (2562.06MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=12|ip-26-0-163-226]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=11|ip-26-0-163-226]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=7|ip-26-0-163-147]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=3|ip-26-0-163-147]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=6|ip-26-0-163-147]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=2|ip-26-0-163-147]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=5|ip-26-0-163-147]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=4|ip-26-0-163-147]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=8|ip-26-0-163-226]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=1|ip-26-0-163-147]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=14|ip-26-0-163-226]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=9|ip-26-0-163-226]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=10|ip-26-0-163-226]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=13|ip-26-0-163-226]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=15|ip-26-0-163-226]: Local number of parameters: 84M (160.13MiB) +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=12|ip-26-0-163-226]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=4|ip-26-0-163-147]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=7|ip-26-0-163-147]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=5|ip-26-0-163-147]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=11|ip-26-0-163-226]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=3|ip-26-0-163-147]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=2|ip-26-0-163-147]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=6|ip-26-0-163-147]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=14|ip-26-0-163-226]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=8|ip-26-0-163-226]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=9|ip-26-0-163-226]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=10|ip-26-0-163-226]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=1|ip-26-0-163-147]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=13|ip-26-0-163-226]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:56 [INFO|DP=0|PP=0|TP=15|ip-26-0-163-226]: [After model building] Memory usage: 160.15MiB. Peak allocated: 5392.00MiB Peak reserved: 22814.00MiB +12/19/2024 02:04:58 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: No checkpoint path provided. +12/19/2024 02:04:58 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Parametrizing model parameters using StandardParametrizator +12/19/2024 02:04:58 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: [Optimizer Building] Using LearningRateForSP as learning rate +12/19/2024 02:04:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/19/2024 02:04:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Using dummy data generator +12/19/2024 02:04:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: [Training Plan] There are 1 training stages +12/19/2024 02:04:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: [Stage Stable Training Stage] start from step 1 +12/19/2024 02:04:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: +12/19/2024 02:04:59 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: [Start training] datetime: 2024-12-19 02:04:59.019993 | mbs: 16 | grad_accum: 2 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +12/19/2024 02:05:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/19/2024 02:05:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Memory usage: 960.80MiB. Peak allocated 5392.00MiB. Peak reserved: 22814.00MiB +12/19/2024 02:05:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Memory usage: 1545.92MiB. Peak allocated 12341.84MiB. Peak reserved: 19586.00MiB +12/19/2024 02:05:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 2.9K | tokens_per_sec: 362K | tokens_per_sec_per_gpu: 2.83K | global_batch_size: 256 | lm_loss: 12.2 | lr: 0.00015 | model_tflops_per_gpu: 27.3 | hardware_tflops_per_gpu: 27.3 | grad_norm: 0.393 | cuda_memory_allocated: 2.29G | cuda_max_memory_reserved: 20.6G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.8G | hd_free_memory_tb: 242G +12/19/2024 02:05:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Memory usage: 2186.49MiB. Peak allocated 2186.49MiB. Peak reserved: 19612.00MiB +12/19/2024 02:05:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Memory usage: 2186.49MiB. Peak allocated 12982.66MiB. Peak reserved: 26932.00MiB +12/19/2024 02:05:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 786 | tokens_per_sec: 1.33M | tokens_per_sec_per_gpu: 10.4K | global_batch_size: 256 | lm_loss: 12.2 | lr: 0.0003 | model_tflops_per_gpu: 101 | hardware_tflops_per_gpu: 101 | grad_norm: 0.393 | cuda_memory_allocated: 2.29G | cuda_max_memory_reserved: 28.2G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.8G | hd_free_memory_tb: 242G +12/19/2024 02:05:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Memory usage: 2186.49MiB. Peak allocated 2186.54MiB. Peak reserved: 26932.00MiB +12/19/2024 02:05:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Memory usage: 2186.49MiB. Peak allocated 12982.66MiB. Peak reserved: 26932.00MiB +12/19/2024 02:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 789 | tokens_per_sec: 1.33M | tokens_per_sec_per_gpu: 10.4K | global_batch_size: 256 | lm_loss: 12.2 | lr: 0.000296 | model_tflops_per_gpu: 100 | hardware_tflops_per_gpu: 100 | grad_norm: 0.385 | cuda_memory_allocated: 2.29G | cuda_max_memory_reserved: 28.2G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.8G | hd_free_memory_tb: 242G +12/19/2024 02:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/19/2024 02:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: | -------- | --------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | ----------------------------------- | -------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/19/2024 02:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: | 13502136 | 1.34G_dp8_tp16_pp1_acc2_mbs16_seq4096_zero0_tpmodeRED_vocab131k | 16 | 4096 | 16 | 2 | 256 | 100.30 | 100.30 | 10378.65 | 199.44 | 124.50 | 125.38 | 458.33 | 260.56 | 262.12 | 2.14 | 26.30 | 8 | 1 | 16 | OneForwardOneBackwardPipelineEngine | REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 32 | 4096 | 131072 | True | torch.bfloat16 | 0 | 25 | True | 1.34G | 84M | +12/19/2024 02:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final2.csv +12/19/2024 02:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Acquired lock for CSV file: benchmark/results/bench_final2.csv +12/19/2024 02:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Successfully wrote to CSV file: benchmark/results/bench_final2.csv. Releasing lock... +12/19/2024 02:05:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-147]: Throughput logging complete +[2024-12-19 02:05:26,610] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 429074 closing signal SIGTERM +[2024-12-19 02:05:26,610] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 429075 closing signal SIGTERM +[2024-12-19 02:05:26,610] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 429076 closing signal SIGTERM +[2024-12-19 02:05:26,610] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 429077 closing signal SIGTERM +[2024-12-19 02:05:26,610] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 429078 closing signal SIGTERM +[2024-12-19 02:05:26,610] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 429079 closing signal SIGTERM +[2024-12-19 02:05:26,615] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 434862 closing signal SIGTERM +[2024-12-19 02:05:26,615] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 434864 closing signal SIGTERM +[2024-12-19 02:05:26,615] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 434865 closing signal SIGTERM +[2024-12-19 02:05:26,615] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 434866 closing signal SIGTERM +[2024-12-19 02:05:26,615] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 434867 closing signal SIGTERM +[2024-12-19 02:05:28,396] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 434860) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-19_02:05:26 + host : ip-26-0-163-147.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 434861) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-19_02:05:26 + host : ip-26-0-163-147.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 434863) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-19_02:05:26 + host : ip-26-0-163-147.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 434860) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-163-147: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13502136.0 +[2024-12-19 02:05:28,812] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,813] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77079 closing signal SIGTERM +[2024-12-19 02:05:28,812] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 503101 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77080 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77081 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 503103 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77082 closing signal SIGTERM +[2024-12-19 02:05:28,811] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,813] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,813] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 59091 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 503104 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77083 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77084 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 59092 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 77085 closing signal SIGTERM +[2024-12-19 02:05:28,812] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711200 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28507 closing signal SIGTERM +[2024-12-19 02:05:28,812] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711201 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 503105 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 503106 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 503107 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 59094 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 458053 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 595267 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28508 closing signal SIGTERM +[2024-12-19 02:05:28,812] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711202 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 458054 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28509 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 59095 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 59096 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28510 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28511 closing signal SIGTERM +[2024-12-19 02:05:28,812] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711204 closing signal SIGTERM +[2024-12-19 02:05:28,812] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711205 closing signal SIGTERM +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 458055 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 59097 closing signal SIGTERM +[2024-12-19 02:05:28,812] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711206 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 595268 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 28513 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 429076 closing signal SIGTERM +[2024-12-19 02:05:28,812] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711207 closing signal SIGTERM +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 458056 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 429077 closing signal SIGTERM +[2024-12-19 02:05:28,809] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 458058 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 595269 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2634409 closing signal SIGTERM +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2634410 closing signal SIGTERM +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2634411 closing signal SIGTERM +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 458059 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 595270 closing signal SIGTERM +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2634412 closing signal SIGTERM +[2024-12-19 02:05:28,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193157 closing signal SIGTERM +[2024-12-19 02:05:28,812] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2634413 closing signal SIGTERM +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 458060 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 595271 closing signal SIGTERM +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2634414 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193159 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 595272 closing signal SIGTERM +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2634415 closing signal SIGTERM +[2024-12-19 02:05:28,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2634416 closing signal SIGTERM +[2024-12-19 02:05:28,812] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 248287 closing signal SIGTERM +[2024-12-19 02:05:28,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193160 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193351 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193352 closing signal SIGTERM +[2024-12-19 02:05:28,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193161 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193353 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193354 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,812] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 248288 closing signal SIGTERM +[2024-12-19 02:05:28,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193162 closing signal SIGTERM +[2024-12-19 02:05:28,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193163 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193355 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193356 closing signal SIGTERM +[2024-12-19 02:05:28,812] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 248289 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 446305 closing signal SIGTERM +[2024-12-19 02:05:28,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193164 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 193357 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 248290 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 446306 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140616 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 446307 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 248291 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140617 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 446309 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140619 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 35563 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 35564 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 446310 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140620 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 35565 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 446311 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 35566 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 446312 closing signal SIGTERM +[2024-12-19 02:05:28,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140621 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 35567 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140622 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 35568 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140623 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 35569 closing signal SIGTERM +[2024-12-19 02:05:28,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 35570 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 429003 got signal: 15 +srun: error: ip-26-0-175-34: task 13: Exited with exit code 1 +[2024-12-19 02:05:30,640] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-174-100.ec2.internal_59017_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 503014 got signal: 15 +srun: error: ip-26-0-174-186: task 10: Exited with exit code 1 +[2024-12-19 02:05:31,382] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-52.ec2.internal_457983_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +[2024-12-19 02:05:31,389] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-175-165.ec2.internal_248212_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +[2024-12-19 02:05:31,392] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-38.ec2.internal_193086_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +[2024-12-19 02:05:31,400] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-24.ec2.internal_76993_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +[2024-12-19 02:05:31,407] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-236.ec2.internal_193280_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 28434 got signal: 15 +[2024-12-19 02:05:31,478] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-174-240.ec2.internal_140545_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 248212 got signal: 15 +[2024-12-19 02:05:31,516] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-18.ec2.internal_446232_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +[2024-12-19 02:05:31,517] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-174-196.ec2.internal_35490_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +[2024-12-19 02:05:31,554] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-142.ec2.internal_2634337_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +[2024-12-19 02:05:31,567] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-175-132.ec2.internal_595180_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +[2024-12-19 02:05:31,574] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-88.ec2.internal_711127_0' has failed to send a keep-alive heartbeat to the rendezvous '13502136' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-175-165: task 15: Exited with exit code 1 +srun: error: ip-26-0-163-226: task 1: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 59017 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 76993 got signal: 15 +srun: error: ip-26-0-174-100: task 9: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 595180 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 446232 got signal: 15 +srun: error: ip-26-0-165-24: task 4: Exited with exit code 1 +srun: error: ip-26-0-175-132: task 14: Exited with exit code 1 +srun: error: ip-26-0-164-18: task 2: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 711127 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 457983 got signal: 15 +srun: error: ip-26-0-171-88: task 7: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 193280 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 35490 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 193086 got signal: 15 +srun: error: ip-26-0-168-52: task 6: Exited with exit code 1 +srun: error: ip-26-0-164-236: task 3: Exited with exit code 1 +srun: error: ip-26-0-174-196: task 11: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2634337 got signal: 15 +srun: error: ip-26-0-165-38: task 5: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 140545 got signal: 15 +srun: error: ip-26-0-172-142: task 8: Exited with exit code 1 +srun: error: ip-26-0-174-240: task 12: Exited with exit code 1 diff --git a/logs/13505355-bench_1.34G_dp1_tp32_pp1_acc1_mbs256_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/13505355-bench_1.34G_dp1_tp32_pp1_acc1_mbs256_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..72cef93f880ae63fc6bb18d0ecc633ac0bb6842b --- /dev/null +++ b/logs/13505355-bench_1.34G_dp1_tp32_pp1_acc1_mbs256_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,2170 @@ ++ '[' -z 13505355 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/slurm/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/opt/slurm/bin:1:/fsx/nouamane/miniconda/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/opt/slurm/bin:1:/fsx/nouamane/miniconda/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-171-[168,230,249],ip-26-0-172-57' ++ export 'NODELIST=ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57' ++ NODELIST='ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-171-[168,230,249],ip-26-0-172-57' ++ export MASTER_NODE=ip-26-0-171-168 ++ MASTER_NODE=ip-26-0-171-168 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-171-168' +Master node: ip-26-0-171-168 ++ echo 'All nodes: ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57' +All nodes: ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 ++ echo 'World size: 32' +World size: 32 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13505355 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-171-168:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.34G_dp1_tp32_pp1_acc1_mbs256_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-19 11:25:34,383] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 11:25:34,383] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 11:25:34,384] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 11:25:34,413] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-19 11:25:34,383] torch.distributed.run: [WARNING] +[2024-12-19 11:25:34,383] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 11:25:34,383] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 11:25:34,383] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 11:25:34,384] torch.distributed.run: [WARNING] +[2024-12-19 11:25:34,384] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 11:25:34,384] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 11:25:34,384] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 11:25:34,384] torch.distributed.run: [WARNING] +[2024-12-19 11:25:34,384] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 11:25:34,384] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 11:25:34,384] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 11:25:34,413] torch.distributed.run: [WARNING] +[2024-12-19 11:25:34,413] torch.distributed.run: [WARNING] ***************************************** +[2024-12-19 11:25:34,413] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-19 11:25:34,413] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +12/19/2024 11:25:58 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Measuring inter-GPU and intra-node bandwidth... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Bandwidth measurement complete. Time taken: 16.56 seconds +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Config: +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Config(general=GeneralArgs(project='debug', +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: run='1.34G_dp1_tp32_pp1_acc1_mbs256_seq4096_zero0_tpmodeRED_vocab131k', +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: seed=42, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: step=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: consumed_train_samples=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: ignore_sanity_checks=True), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: parallelism=ParallelismArgs(dp=1, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pp=1, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tp=32, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pp_engine=, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tp_mode=, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tp_linear_async_communication=True, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: recompute_layer=False, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tp_recompute_allgather=True, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: expert_parallel_size=1), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: eos_token_id=0, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: hidden_act='silu', +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: hidden_size=2048, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: initializer_range=0.02, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: intermediate_size=8192, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: is_llama_config=True, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: max_position_embeddings=4096, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_attention_heads=32, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_hidden_layers=16, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_key_value_heads=32, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pad_token_id=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pretraining_tp=1, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rms_norm_eps=1e-05, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_scaling=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_theta=10000.0, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_interleaved=False, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tie_word_embeddings=True, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: use_cache=True, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: vocab_size=131072), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: init_method=RandomInit(std=0.02), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: dtype=torch.bfloat16, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: make_vocab_size_divisible_by=1, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: ddp_bucket_cap_mb=25), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tokenizer_revision=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tokenizer_max_length=None), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: checkpoint_interval=10000, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: save_initial_state=False, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: save_final_state=False, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: resume_checkpoint_path=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: checkpoints_path_is_shared_file_system=False), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: logging=LoggingArgs(log_level='info', +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: log_level_replica='info', +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: iteration_step_info_interval=1), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tokens=TokensArgs(sequence_length=4096, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: train_steps=100, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: micro_batch_size=256, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: batch_accumulation_per_replica=1, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: val_check_interval=100, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: limit_val_batches=0, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: limit_test_batches=0), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: adam_beta1=0.9, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: adam_beta2=0.95, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: torch_adam_is_fused=True, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: name='adamW'), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: zero_stage=0, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: weight_decay=0.01, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: clip_grad=1.0, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: accumulate_grad_in_fp32=True, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lr_warmup_steps=2, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lr_warmup_style='linear', +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lr_decay_style='cosine', +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lr_decay_steps=13, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lr_decay_starting_step=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: min_decay_lr=1e-05)), +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: start_training_step=1, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: data=DataArgs(dataset=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: seed=42, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_loading_workers=1))], +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: profiler=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: lighteval=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: s3_upload=None) +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Model Config: +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: LlamaConfig(bos_token_id=0, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: eos_token_id=0, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: hidden_act='silu', +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: hidden_size=2048, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: initializer_range=0.02, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: intermediate_size=8192, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: is_llama_config=True, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: max_position_embeddings=4096, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_attention_heads=32, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_hidden_layers=16, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: num_key_value_heads=32, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pad_token_id=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: pretraining_tp=1, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rms_norm_eps=1e-05, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_scaling=None, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_theta=10000.0, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: rope_interleaved=False, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: tie_word_embeddings=True, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: use_cache=True, +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: vocab_size=131072) +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Building model.. +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Initialize RoPE Theta = 10000.0 +12/19/2024 11:26:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=23|ip-26-0-171-249]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=17|ip-26-0-171-249]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=6|ip-26-0-171-168]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Total number of parameters: 1.34G (2564.12MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=7|ip-26-0-171-168]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=4|ip-26-0-171-168]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=18|ip-26-0-171-249]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=19|ip-26-0-171-249]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=14|ip-26-0-171-230]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=13|ip-26-0-171-230]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=9|ip-26-0-171-230]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=10|ip-26-0-171-230]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=16|ip-26-0-171-249]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=21|ip-26-0-171-249]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=20|ip-26-0-171-249]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=22|ip-26-0-171-249]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=2|ip-26-0-171-168]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=5|ip-26-0-171-168]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=1|ip-26-0-171-168]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=15|ip-26-0-171-230]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=8|ip-26-0-171-230]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=11|ip-26-0-171-230]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=12|ip-26-0-171-230]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=3|ip-26-0-171-168]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=23|ip-26-0-171-249]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=7|ip-26-0-171-168]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 20834.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=6|ip-26-0-171-168]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 18786.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=17|ip-26-0-171-249]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=4|ip-26-0-171-168]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 18786.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=18|ip-26-0-171-249]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=16|ip-26-0-171-249]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=19|ip-26-0-171-249]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=21|ip-26-0-171-249]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=20|ip-26-0-171-249]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=13|ip-26-0-171-230]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=14|ip-26-0-171-230]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=10|ip-26-0-171-230]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 18786.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=29|ip-26-0-172-57]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=5|ip-26-0-171-168]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 18786.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=2|ip-26-0-171-168]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 20834.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=15|ip-26-0-171-230]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=28|ip-26-0-172-57]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=25|ip-26-0-172-57]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=9|ip-26-0-171-230]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=22|ip-26-0-171-249]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=11|ip-26-0-171-230]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=8|ip-26-0-171-230]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=1|ip-26-0-171-168]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 19810.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=26|ip-26-0-172-57]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=30|ip-26-0-172-57]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=3|ip-26-0-171-168]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 20834.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=24|ip-26-0-172-57]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=12|ip-26-0-171-230]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=29|ip-26-0-172-57]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=28|ip-26-0-172-57]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=25|ip-26-0-172-57]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=26|ip-26-0-172-57]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=30|ip-26-0-172-57]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=24|ip-26-0-172-57]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: No checkpoint path provided. +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Parametrizing model parameters using StandardParametrizator +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=27|ip-26-0-172-57]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=31|ip-26-0-172-57]: Local number of parameters: 42M (80.13MiB) +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=27|ip-26-0-172-57]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=31|ip-26-0-172-57]: [After model building] Memory usage: 80.15MiB. Peak allocated: 5440.00MiB Peak reserved: 22882.00MiB +12/19/2024 11:26:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [Optimizer Building] Using LearningRateForSP as learning rate +12/19/2024 11:26:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/19/2024 11:26:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Using dummy data generator +12/19/2024 11:26:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [Training Plan] There are 1 training stages +12/19/2024 11:26:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [Stage Stable Training Stage] start from step 1 +12/19/2024 11:26:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: +12/19/2024 11:26:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: [Start training] datetime: 2024-12-19 11:26:26.148095 | mbs: 256 | grad_accum: 1 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +12/19/2024 11:26:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/19/2024 11:26:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-171-168]: Memory usage: 400.67MiB. Peak allocated 5440.00MiB. Peak reserved: 18786.00MiB +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs)return f(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + output = model(**micro_batch) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model(sharded_logits = self.model( + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +sharded_logits = self.model( File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + outputs = self.pipeline_engine.train_batch_iter( +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter(return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in +output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in +output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 3.43 GiB is free. Including non-PyTorch memory, this process has 75.89 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 474.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + module_builder=lambda: lambda x: x.float(), +module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 3.28 GiB is free. Including non-PyTorch memory, this process has 76.04 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 602.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFtorch.cuda +.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 3.40 GiB is free. Including non-PyTorch memory, this process has 75.92 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 474.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + module_builder=lambda: lambda x: x.float(), + return forward_call(*args, **kwargs)torch.cuda +.OutOfMemoryError File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 3.53 GiB is free. Including non-PyTorch memory, this process has 75.79 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 346.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 3.38 GiB is free. Including non-PyTorch memory, this process has 75.94 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 474.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 3.45 GiB is free. Including non-PyTorch memory, this process has 75.87 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 602.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 3.23 GiB is free. Including non-PyTorch memory, this process has 76.09 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 474.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 1.53 GiB is free. Including non-PyTorch memory, this process has 77.79 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 2.34 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch)trainer.train(dataloader) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) trainer.train(dataloader) +trainer.train(dataloader) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +sharded_logits = self.model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter +output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs = self.pipeline_engine.train_batch_iter( +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return forward_call(*args, **kwargs) + +return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward +return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model(sharded_logits = self.model(return forward_call(*args, **kwargs) + + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model(sharded_logits = self.model( + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + output = self.pp_block(**new_kwargs)sharded_logits = self.model( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + sharded_logits = self.model( +module_builder=lambda: lambda x: x.float(), + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 2.40 GiB is free. Including non-PyTorch memory, this process has 76.92 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 602.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 979.94 MiB is free. Including non-PyTorch memory, this process has 78.36 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 2.09 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] +fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + output = self.pp_block(**new_kwargs)output = self.pp_block(**new_kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), + module_builder=lambda: lambda x: x.float(), + torch.cudamodule_builder=lambda: lambda x: x.float(),. +OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 2.65 GiB is free. Including non-PyTorch memory, this process has 76.67 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 346.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cudatorch.cuda..OutOfMemoryErrorOutOfMemoryError: : CUDA out of memory. Tried to allocate 16.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 619.94 MiB is free. Including non-PyTorch memory, this process has 78.71 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 2.09 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFCUDA out of memory. Tried to allocate 16.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 543.94 MiB is free. Including non-PyTorch memory, this process has 78.79 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 2.34 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 639.94 MiB is free. Including non-PyTorch memory, this process has 78.69 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 2.34 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 2.42 GiB is free. Including non-PyTorch memory, this process has 76.90 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 602.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 2.53 GiB is free. Including non-PyTorch memory, this process has 76.79 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 346.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 3.55 GiB is free. Including non-PyTorch memory, this process has 75.77 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 346.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader)trainer.train(dataloader) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter(outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) +output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] +sharded_logits = self.model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = model(**micro_batch)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states +output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + module_builder=lambda: lambda x: x.float(), + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 3.28 GiB is free. Including non-PyTorch memory, this process has 76.04 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 602.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = self.pp_block(**new_kwargs) +output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + module_builder=lambda: lambda x: x.float(), + torch.cudareturn self._call_impl(*args, **kwargs). +OutOfMemoryError: File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +CUDA out of memory. Tried to allocate 16.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 3.45 GiB is free. Including non-PyTorch memory, this process has 75.87 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 602.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 3.15 GiB is free. Including non-PyTorch memory, this process has 76.17 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 730.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + module_builder=lambda: lambda x: x.float(), File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 3.38 GiB is free. Including non-PyTorch memory, this process has 75.94 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 474.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 3.53 GiB is free. Including non-PyTorch memory, this process has 75.79 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 346.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 3.53 GiB is free. Including non-PyTorch memory, this process has 75.79 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 346.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + return f(*args, **kwargs) + module_builder=lambda: lambda x: x.float(), + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 3.10 GiB is free. Including non-PyTorch memory, this process has 76.22 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 602.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 907, in forward_with_hidden_states +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + sharded_logits = self.lm_head(x=hidden_states)["logits"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 179, in forward +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] +trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + gathered_output = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 8.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 6.80 GiB is free. Including non-PyTorch memory, this process has 72.51 GiB memory in use. Of the allocated memory 56.77 GiB is allocated by PyTorch, and 4.09 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return f(*args, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 415.94 MiB is free. Including non-PyTorch memory, this process has 78.91 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 2.46 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 391.94 MiB is free. Including non-PyTorch memory, this process has 78.94 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 2.46 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 2.22 GiB is free. Including non-PyTorch memory, this process has 77.10 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 602.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + return forward_call(*args, **kwargs)output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model)return f(*args, **kwargs) + + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + sharded_logits = self.model( File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) +output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 2.40 GiB is free. Including non-PyTorch memory, this process has 76.92 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 602.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 843.94 MiB is free. Including non-PyTorch memory, this process has 78.49 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 2.34 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 2.40 GiB is free. Including non-PyTorch memory, this process has 76.92 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 602.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 543.94 MiB is free. Including non-PyTorch memory, this process has 78.79 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 2.34 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[2024-12-19 11:26:56,730] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19855 closing signal SIGTERM +[2024-12-19 11:26:56,730] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19856 closing signal SIGTERM +[2024-12-19 11:26:56,730] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19858 closing signal SIGTERM +[2024-12-19 11:26:56,730] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19859 closing signal SIGTERM +[2024-12-19 11:26:56,730] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19861 closing signal SIGTERM +[2024-12-19 11:26:56,735] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447726 closing signal SIGTERM +[2024-12-19 11:26:56,736] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447727 closing signal SIGTERM +[2024-12-19 11:26:56,736] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447728 closing signal SIGTERM +[2024-12-19 11:26:56,736] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447729 closing signal SIGTERM +[2024-12-19 11:26:56,736] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447731 closing signal SIGTERM +[2024-12-19 11:26:56,736] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447732 closing signal SIGTERM +[2024-12-19 11:26:58,323] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 19854) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-19 11:26:58,377] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_3w_5ocgy/13505355_1oimoy7y/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-19_11:26:37 + host : ip-26-0-172-57.ec2.internal + rank : 27 (local_rank: 3) + exitcode : 1 (pid: 19857) + error_file: /tmp/torchelastic_3w_5ocgy/13505355_1oimoy7y/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 3.23 GiB is free. Including non-PyTorch memory, this process has 76.09 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 474.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[2]: + time : 2024-12-19_11:26:37 + host : ip-26-0-172-57.ec2.internal + rank : 30 (local_rank: 6) + exitcode : 1 (pid: 19860) + error_file: /tmp/torchelastic_3w_5ocgy/13505355_1oimoy7y/attempt_0/6/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 3.28 GiB is free. Including non-PyTorch memory, this process has 76.04 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 602.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-19_11:26:37 + host : ip-26-0-172-57.ec2.internal + rank : 24 (local_rank: 0) + exitcode : 1 (pid: 19854) + error_file: /tmp/torchelastic_3w_5ocgy/13505355_1oimoy7y/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 50, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1013, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 151, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 3.43 GiB is free. Including non-PyTorch memory, this process has 75.89 GiB memory in use. Of the allocated memory 64.77 GiB is allocated by PyTorch, and 474.15 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +============================================================ +srun: error: ip-26-0-172-57: task 3: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13505355.0 +slurmstepd: error: *** STEP 13505355.0 ON ip-26-0-171-168 CANCELLED AT 2024-12-19T11:26:58 *** +[2024-12-19 11:26:58,737] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 11:26:58,737] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1447728 closing signal SIGTERM +[2024-12-19 11:26:58,737] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 11:26:58,737] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 455851 closing signal SIGTERM +[2024-12-19 11:26:58,737] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 455852 closing signal SIGTERM +[2024-12-19 11:26:58,737] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 455853 closing signal SIGTERM +[2024-12-19 11:26:58,737] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 455854 closing signal SIGTERM +[2024-12-19 11:26:58,737] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 455855 closing signal SIGTERM +[2024-12-19 11:26:58,737] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 455856 closing signal SIGTERM +[2024-12-19 11:26:58,737] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-19 11:26:58,737] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 455857 closing signal SIGTERM +[2024-12-19 11:26:58,737] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53903 closing signal SIGTERM +[2024-12-19 11:26:58,737] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 455858 closing signal SIGTERM +[2024-12-19 11:26:58,737] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53904 closing signal SIGTERM +[2024-12-19 11:26:58,738] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53905 closing signal SIGTERM +[2024-12-19 11:26:58,738] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53906 closing signal SIGTERM +[2024-12-19 11:26:58,738] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53907 closing signal SIGTERM +[2024-12-19 11:26:58,738] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53908 closing signal SIGTERM +[2024-12-19 11:26:58,738] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53909 closing signal SIGTERM +[2024-12-19 11:26:58,738] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53910 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1447650 got signal: 15 +srun: error: ip-26-0-171-168: task 0: Exited with exit code 1 +[2024-12-19 11:27:01,512] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-249.ec2.internal_53832_0' has failed to send a keep-alive heartbeat to the rendezvous '13505355' due to an error of type RendezvousConnectionError. +[2024-12-19 11:27:01,626] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-230.ec2.internal_455780_0' has failed to send a keep-alive heartbeat to the rendezvous '13505355' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 53832 got signal: 15 +srun: error: ip-26-0-171-249: task 2: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 455780 got signal: 15 +srun: error: ip-26-0-171-230: task 1: Exited with exit code 1 diff --git a/logs/13523084-bench.out b/logs/13523084-bench.out new file mode 100644 index 0000000000000000000000000000000000000000..58143c47337a2c2fca38f1ed4766117bb76f2779 --- /dev/null +++ b/logs/13523084-bench.out @@ -0,0 +1,5638 @@ ++ '[' -z 13523084 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/slurm/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/opt/slurm/bin:1:/fsx/nouamane/miniconda/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/opt/slurm/bin:1:/fsx/nouamane/miniconda/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-174-[100,186]' ++ export 'NODELIST=ip-26-0-174-100 +ip-26-0-174-186' ++ NODELIST='ip-26-0-174-100 +ip-26-0-174-186' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-174-[100,186]' ++ export MASTER_NODE=ip-26-0-174-100 ++ MASTER_NODE=ip-26-0-174-100 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export NCCL_NET_GDR_LEVEL=LOC ++ NCCL_NET_GDR_LEVEL=LOC ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-43 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 52-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 52-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 52-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 52-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-174-100' +Master node: ip-26-0-174-100 ++ echo 'All nodes: ip-26-0-174-100 +ip-26-0-174-186' +All nodes: ip-26-0-174-100 +ip-26-0-174-186 ++ echo 'World size: 16' +World size: 16 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13523084 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-174-100:12356 --max_restarts 0 --rdzv_conf timeout=60 test_pp_8gpus.py +[2024-12-20 12:01:54,094] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-20 12:01:54,095] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-20 12:01:54,094] torch.distributed.run: [WARNING] +[2024-12-20 12:01:54,094] torch.distributed.run: [WARNING] ***************************************** +[2024-12-20 12:01:54,094] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-20 12:01:54,094] torch.distributed.run: [WARNING] ***************************************** +[2024-12-20 12:01:54,095] torch.distributed.run: [WARNING] +[2024-12-20 12:01:54,095] torch.distributed.run: [WARNING] ***************************************** +[2024-12-20 12:01:54,095] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-20 12:01:54,095] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-174-100:70019:70019 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-100:70019:70019 [0] NCCL INFO Bootstrap : Using enp72s0:26.0.172.74<0> +ip-26-0-174-100:70019:70019 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-100:70019:70019 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-100:70019:70019 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-100:70025:70025 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-100:70025:70025 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-100:70023:70023 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-100:70026:70026 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-100:70020:70020 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-100:70022:70022 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-100:70024:70024 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-100:70021:70021 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-100:70023:70023 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-100:70026:70026 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-100:70020:70020 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-100:70022:70022 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-100:70024:70024 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-100:70021:70021 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-100:70025:70025 [6] NCCL INFO Bootstrap : Using enp72s0:26.0.172.74<0> +ip-26-0-174-186:375444:375444 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-186:375439:375439 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-100:70025:70025 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-100:70025:70025 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-186:375445:375445 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-186:375439:375439 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-186:375444:375444 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-186:375445:375445 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-100:70023:70023 [4] NCCL INFO Bootstrap : Using enp72s0:26.0.172.74<0> +ip-26-0-174-100:70026:70026 [7] NCCL INFO Bootstrap : Using enp72s0:26.0.172.74<0> +ip-26-0-174-100:70020:70020 [1] NCCL INFO Bootstrap : Using enp72s0:26.0.172.74<0> +ip-26-0-174-100:70021:70021 [2] NCCL INFO Bootstrap : Using enp72s0:26.0.172.74<0> +ip-26-0-174-100:70022:70022 [3] NCCL INFO Bootstrap : Using enp72s0:26.0.172.74<0> +ip-26-0-174-100:70024:70024 [5] NCCL INFO Bootstrap : Using enp72s0:26.0.172.74<0> +ip-26-0-174-100:70026:70026 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-100:70026:70026 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-100:70023:70023 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-100:70024:70024 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-100:70020:70020 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-100:70023:70023 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-100:70024:70024 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-100:70020:70020 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-100:70022:70022 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-100:70021:70021 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-100:70021:70021 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-100:70022:70022 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-186:375444:375444 [6] NCCL INFO Bootstrap : Using enp72s0:26.0.175.189<0> +ip-26-0-174-186:375439:375439 [1] NCCL INFO Bootstrap : Using enp72s0:26.0.175.189<0> +ip-26-0-174-186:375445:375445 [7] NCCL INFO Bootstrap : Using enp72s0:26.0.175.189<0> +ip-26-0-174-186:375443:375443 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-186:375443:375443 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-186:375441:375441 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-186:375442:375442 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-186:375440:375440 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-186:375441:375441 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-186:375442:375442 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-186:375440:375440 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-186:375439:375439 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-186:375444:375444 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-186:375439:375439 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-186:375444:375444 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-186:375445:375445 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-186:375445:375445 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-186:375443:375443 [5] NCCL INFO Bootstrap : Using enp72s0:26.0.175.189<0> +ip-26-0-174-186:375443:375443 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-186:375443:375443 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-186:375441:375441 [3] NCCL INFO Bootstrap : Using enp72s0:26.0.175.189<0> +ip-26-0-174-186:375442:375442 [4] NCCL INFO Bootstrap : Using enp72s0:26.0.175.189<0> +ip-26-0-174-186:375440:375440 [2] NCCL INFO Bootstrap : Using enp72s0:26.0.175.189<0> +ip-26-0-174-186:375441:375441 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-186:375441:375441 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-186:375440:375440 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-186:375440:375440 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-186:375442:375442 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-186:375442:375442 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-186:375438:375438 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-186:375438:375438 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-174-186:375438:375438 [0] NCCL INFO Bootstrap : Using enp72s0:26.0.175.189<0> +ip-26-0-174-186:375438:375438 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-186:375438:375438 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-100:70024:70101 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-100:70024:70101 [5] NCCL INFO Using network Libfabric +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/167 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/146 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-100:70026:70097 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-100:70026:70097 [7] NCCL INFO Using network Libfabric +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-100:70022:70096 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-100:70022:70096 [3] NCCL INFO Using network Libfabric +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-100:70023:70098 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-100:70023:70098 [4] NCCL INFO Using network Libfabric +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-100:70025:70095 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-100:70025:70095 [6] NCCL INFO Using network Libfabric +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-100:70021:70100 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-100:70021:70100 [2] NCCL INFO Using network Libfabric +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-100:70020:70099 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-100:70020:70099 [1] NCCL INFO Using network Libfabric +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-100:70019:70094 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-100:70019:70094 [0] NCCL INFO Using network Libfabric +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-186:375439:375515 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-186:375439:375515 [1] NCCL INFO Using network Libfabric +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-186:375441:375511 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-186:375441:375511 [3] NCCL INFO Using network Libfabric +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-186:375442:375514 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-186:375442:375514 [4] NCCL INFO Using network Libfabric +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-186:375440:375513 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-186:375440:375513 [2] NCCL INFO Using network Libfabric +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-186:375443:375512 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-186:375443:375512 [5] NCCL INFO Using network Libfabric +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-186:375445:375517 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-186:375445:375517 [7] NCCL INFO Using network Libfabric +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-186:375444:375516 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-186:375444:375516 [6] NCCL INFO Using network Libfabric +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-186:375438:375518 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-186:375438:375518 [0] NCCL INFO Using network Libfabric +ip-26-0-174-186:375444:375516 [6] NCCL INFO comm 0x8397040 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-186:375445:375517 [7] NCCL INFO comm 0x8b667b0 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-100:70020:70099 [1] NCCL INFO comm 0x89ba7a0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-100:70023:70098 [4] NCCL INFO comm 0x7c1af20 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-100:70021:70100 [2] NCCL INFO comm 0x7f291a0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-100:70022:70096 [3] NCCL INFO comm 0x7931080 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-100:70019:70094 [0] NCCL INFO comm 0x8ee4bc0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-100:70024:70101 [5] NCCL INFO comm 0x84418b0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-186:375443:375512 [5] NCCL INFO comm 0x7220020 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-186:375442:375514 [4] NCCL INFO comm 0x7c1da70 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-100:70025:70095 [6] NCCL INFO comm 0x9010ee0 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-100:70026:70097 [7] NCCL INFO comm 0x8cb6e20 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-186:375438:375518 [0] NCCL INFO comm 0x7668500 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-186:375440:375513 [2] NCCL INFO comm 0x7e24920 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-186:375441:375511 [3] NCCL INFO comm 0x79cca40 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-186:375439:375515 [1] NCCL INFO comm 0x74129a0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf6cc9c8537c5a7e4 - Init START +ip-26-0-174-186:375443:375512 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/167 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375441:375511 [3] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-174-186:375441:375511 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-100:70019:70094 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-174-100:70019:70094 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-174-100:70021:70100 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375444:375516 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-174-186:375438:375518 [0] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-174-186:375438:375518 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-174-186:375439:375515 [1] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-174-186:375439:375515 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-174-100:70020:70099 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-186:375445:375517 [7] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-174-186:375440:375513 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-174-186:375445:375517 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375442:375514 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-100:70026:70097 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-174-100:70023:70098 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-174-100:70022:70096 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NCCL_NET_GDR_LEVEL set by environment to LOC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375443:375512 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->1 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->1 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375445:375517 [7] NCCL INFO Trees [0] 8/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->13 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->13 +ip-26-0-174-186:375445:375517 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70023:70098 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] -1/-1/-1->4->3 [3] -1/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] -1/-1/-1->4->3 [7] -1/-1/-1->4->3 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Trees [0] 10/-1/-1->9->1 [1] 11/-1/-1->9->10 [2] 10/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/1/-1->9->-1 [5] 11/-1/-1->9->10 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 +ip-26-0-174-186:375439:375515 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->9 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->10 [4] 12/-1/-1->11->10 [5] 12/-1/-1->11->9 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 +ip-26-0-174-186:375443:375512 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->5 [3] 15/-1/-1->13->14 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->12 [6] 14/5/-1->13->-1 [7] 15/-1/-1->13->14 +ip-26-0-174-186:375444:375516 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 13/-1/-1->14->6 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->13 [7] 13/6/-1->14->-1 +ip-26-0-174-186:375442:375514 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] -1/-1/-1->12->11 [3] -1/-1/-1->12->11 [4] 13/-1/-1->12->11 [5] 13/-1/-1->12->11 [6] -1/-1/-1->12->11 [7] -1/-1/-1->12->11 +ip-26-0-174-186:375441:375511 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Trees [0] -1/-1/-1->8->15 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] -1/-1/-1->8->15 [5] -1/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 +ip-26-0-174-186:375443:375512 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375444:375516 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375442:375514 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375438:375518 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70026:70097 [7] NCCL INFO Trees [0] 0/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->5 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/-1/-1->7->5 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 9/-1/-1->10->2 [2] 11/-1/-1->10->9 [3] 11/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 9/2/-1->10->-1 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70026:70097 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70025:70095 [6] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 00/08 : 0 7 6 5 4 3 2 1 8 15 14 13 12 11 10 9 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 1/10/-1->2->-1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 1/-1/-1->2->10 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 01/08 : 0 7 6 5 4 2 3 9 8 15 14 13 12 10 11 1 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70020:70099 [1] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/13/-1->5->-1 [3] 7/-1/-1->5->6 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->13 [7] 7/-1/-1->5->6 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 02/08 : 0 2 7 6 5 12 11 9 8 10 15 14 13 4 3 1 +ip-26-0-174-100:70025:70095 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 5/14/-1->6->-1 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 5/-1/-1->6->14 +ip-26-0-174-100:70024:70101 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 03/08 : 0 4 2 6 7 13 11 9 8 12 10 14 15 5 3 1 +ip-26-0-174-100:70025:70095 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 04/08 : 0 7 6 5 4 3 2 1 8 15 14 13 12 11 10 9 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 05/08 : 0 7 6 5 4 2 3 9 8 15 14 13 12 10 11 1 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 06/08 : 0 2 7 6 5 12 11 9 8 10 15 14 13 4 3 1 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Trees [0] 2/9/-1->1->-1 [1] 3/-1/-1->1->2 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->9 [5] 3/-1/-1->1->2 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 07/08 : 0 4 2 6 7 13 11 9 8 12 10 14 15 5 3 1 +ip-26-0-174-100:70020:70099 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Trees [0] -1/-1/-1->0->7 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] -1/-1/-1->0->7 [5] -1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 +ip-26-0-174-100:70019:70094 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 02/0 : 8[0] -> 10[2] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 02/0 : 0[0] -> 2[2] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 07/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 06/0 : 8[0] -> 10[2] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 06/0 : 0[0] -> 2[2] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 03/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 03/0 : 10[2] -> 14[6] via P2P/IPC +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 03/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 03/0 : 2[2] -> 6[6] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 07/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 07/0 : 10[2] -> 14[6] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 07/0 : 2[2] -> 6[6] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 01/0 : 11[3] -> 1[1] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 07/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 05/0 : 11[3] -> 1[1] [send] via NET/Libfabric/3 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 02/0 : 10[2] -> 15[7] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 02/0 : 2[2] -> 7[7] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 06/0 : 10[2] -> 15[7] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 01/0 : 3[3] -> 9[1] [send] via NET/Libfabric/3 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 05/0 : 3[3] -> 9[1] [send] via NET/Libfabric/3 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 06/0 : 2[2] -> 7[7] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 03/0 : 15[7] -> 5[5] [send] via NET/Libfabric/7 +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 07/0 : 15[7] -> 5[5] [send] via NET/Libfabric/7 +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 03/0 : 7[7] -> 13[5] [send] via NET/Libfabric/7 +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 07/0 : 7[7] -> 13[5] [send] via NET/Libfabric/7 +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 06/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 03/0 : 15[7] -> 5[5] [receive] via NET/Libfabric/6 +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 07/0 : 15[7] -> 5[5] [receive] via NET/Libfabric/6 +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 02/0 : 5[5] -> 12[4] [send] via NET/Libfabric/6 +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 06/0 : 5[5] -> 12[4] [send] via NET/Libfabric/6 +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 03/0 : 7[7] -> 13[5] [receive] via NET/Libfabric/6 +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 07/0 : 7[7] -> 13[5] [receive] via NET/Libfabric/6 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 01/0 : 3[3] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 05/0 : 3[3] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 00/0 : 9[1] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 02/0 : 13[5] -> 4[4] [send] via NET/Libfabric/6 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 04/0 : 9[1] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 00/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 06/0 : 13[5] -> 4[4] [send] via NET/Libfabric/6 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 00/0 : 9[1] -> 0[0] [send] via NET/Libfabric/2 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 04/0 : 9[1] -> 0[0] [send] via NET/Libfabric/2 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 01/0 : 11[3] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 05/0 : 11[3] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 00/0 : 1[1] -> 8[0] [send] via NET/Libfabric/2 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 04/0 : 1[1] -> 8[0] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 06/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 07/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 00/0 : 1[1] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 02/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4 +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 04/0 : 1[1] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 06/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 00/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 06/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4 +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 07/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 07/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 02/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4 +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 06/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 03/0 : 12[4] -> 10[2] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 05/0 : 12[4] -> 10[2] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 03/0 : 4[4] -> 2[2] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 07/0 : 12[4] -> 10[2] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 05/0 : 4[4] -> 2[2] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 07/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 07/0 : 4[4] -> 2[2] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 05/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Connected all rings +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Connected all rings +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Connected all rings +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Connected all rings +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Connected all rings +ip-26-0-174-100:70019:70094 [0] NCCL INFO Connected all rings +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 05/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Connected all rings +ip-26-0-174-186:375443:375512 [5] NCCL INFO Connected all rings +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 06/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 03/0 : 13[5] -> 15[7] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 07/0 : 13[5] -> 15[7] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Connected all rings +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 04/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-186:375442:375514 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Connected all rings +ip-26-0-174-100:70020:70099 [1] NCCL INFO Connected all rings +ip-26-0-174-100:70021:70100 [2] NCCL INFO Connected all rings +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Connected all rings +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 01/0 : 9[1] -> 11[3] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 05/0 : 9[1] -> 11[3] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Connected all rings +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 01/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 05/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 00/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/4 +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/4 +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [send] via NET/Libfabric/4 +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375443:375512 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [send] via NET/Libfabric/4 +ip-26-0-174-100:70024:70101 [5] NCCL INFO Connected all rings +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/7 +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/7 +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/7 +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/7 +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 03/0 : 15[7] -> 13[5] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-186:375445:375517 [7] NCCL INFO Channel 07/0 : 15[7] -> 13[5] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Connected all rings +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 01/0 : 1[1] -> 3[3] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 05/0 : 1[1] -> 3[3] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 03/0 : 5[5] -> 7[7] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 07/0 : 5[5] -> 7[7] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 01/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 05/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-100:70023:70098 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 03/0 : 7[7] -> 5[5] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Channel 07/0 : 7[7] -> 5[5] via P2P/IPC +ip-26-0-174-186:375438:375518 [0] NCCL INFO Connected all trees +ip-26-0-174-186:375438:375518 [0] NCCL INFO NVLS comm 0x7668500 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Connected all trees +ip-26-0-174-100:70019:70094 [0] NCCL INFO NVLS comm 0x8ee4bc0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [send] via NET/Libfabric/0 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/3 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/3 +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/7 +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/7 +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/7 +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/7 +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-100:70025:70095 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-186:375444:375516 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/4 +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/4 +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [send] via NET/Libfabric/4 +ip-26-0-174-100:70024:70101 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [send] via NET/Libfabric/4 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [send] via NET/Libfabric/0 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/3 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/3 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-100:70026:70097 [7] NCCL INFO Connected all trees +ip-26-0-174-100:70026:70097 [7] NCCL INFO NVLS comm 0x8cb6e20 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375445:375517 [7] NCCL INFO Connected all trees +ip-26-0-174-186:375445:375517 [7] NCCL INFO NVLS comm 0x8b667b0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375442:375514 [4] NCCL INFO Connected all trees +ip-26-0-174-186:375442:375514 [4] NCCL INFO NVLS comm 0x7c1da70 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Connected all trees +ip-26-0-174-186:375441:375511 [3] NCCL INFO NVLS comm 0x79cca40 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70023:70098 [4] NCCL INFO Connected all trees +ip-26-0-174-100:70023:70098 [4] NCCL INFO NVLS comm 0x7c1af20 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375443:375512 [5] NCCL INFO Connected all trees +ip-26-0-174-186:375443:375512 [5] NCCL INFO NVLS comm 0x7220020 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375444:375516 [6] NCCL INFO Connected all trees +ip-26-0-174-186:375444:375516 [6] NCCL INFO NVLS comm 0x8397040 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70025:70095 [6] NCCL INFO Connected all trees +ip-26-0-174-100:70024:70101 [5] NCCL INFO Connected all trees +ip-26-0-174-100:70025:70095 [6] NCCL INFO NVLS comm 0x9010ee0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70024:70101 [5] NCCL INFO NVLS comm 0x84418b0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Connected all trees +ip-26-0-174-100:70022:70096 [3] NCCL INFO NVLS comm 0x7931080 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Connected all trees +ip-26-0-174-186:375439:375515 [1] NCCL INFO NVLS comm 0x74129a0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Connected all trees +ip-26-0-174-186:375440:375513 [2] NCCL INFO NVLS comm 0x7e24920 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Connected all trees +ip-26-0-174-100:70020:70099 [1] NCCL INFO NVLS comm 0x89ba7a0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Connected all trees +ip-26-0-174-100:70021:70100 [2] NCCL INFO NVLS comm 0x7f291a0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375442:375514 [4] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375442:375514 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375442:375514 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375442:375514 [4] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375445:375517 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375445:375517 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375445:375517 [7] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375444:375516 [6] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375444:375516 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375444:375516 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375444:375516 [6] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375443:375512 [5] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375443:375512 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375443:375512 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375443:375512 [5] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375518 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375515 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375441:375511 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70025:70095 [6] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70023:70098 [4] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70025:70095 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70025:70095 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70025:70095 [6] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70023:70098 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70023:70098 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70023:70098 [4] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70024:70101 [5] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70026:70097 [7] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70024:70101 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70024:70101 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70024:70101 [5] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70026:70097 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70026:70097 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70026:70097 [7] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70094 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70022:70096 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375513 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70100 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70020:70099 [1] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70020:70099 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70020:70099 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70020:70099 [1] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375439:375515 [1] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375439:375515 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375439:375515 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375439:375515 [1] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375441:375511 [3] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375441:375511 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375441:375511 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375441:375511 [3] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70019:70094 [0] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70019:70094 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70019:70094 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70019:70094 [0] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70022:70096 [3] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70022:70096 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70022:70096 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70022:70096 [3] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375438:375518 [0] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375438:375518 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375438:375518 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375438:375518 [0] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70021:70100 [2] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70021:70100 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70021:70100 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70021:70100 [2] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70023:70098 [4] NCCL INFO comm 0x7c1af20 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-100:70019:70094 [0] NCCL INFO comm 0x8ee4bc0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-100:70022:70096 [3] NCCL INFO comm 0x7931080 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-100:70021:70100 [2] NCCL INFO comm 0x7f291a0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-100:70025:70095 [6] NCCL INFO comm 0x9010ee0 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-100:70020:70099 [1] NCCL INFO comm 0x89ba7a0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-100:70024:70101 [5] NCCL INFO comm 0x84418b0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-100:70026:70097 [7] NCCL INFO comm 0x8cb6e20 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-186:375440:375513 [2] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375440:375513 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375440:375513 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375440:375513 [2] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375441:375511 [3] NCCL INFO comm 0x79cca40 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-186:375438:375518 [0] NCCL INFO comm 0x7668500 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-186:375440:375513 [2] NCCL INFO comm 0x7e24920 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-186:375445:375517 [7] NCCL INFO comm 0x8b667b0 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-186:375442:375514 [4] NCCL INFO comm 0x7c1da70 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-186:375443:375512 [5] NCCL INFO comm 0x7220020 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-186:375439:375515 [1] NCCL INFO comm 0x74129a0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +ip-26-0-174-186:375444:375516 [6] NCCL INFO comm 0x8397040 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf6cc9c8537c5a7e4 - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-174-100:70019:70250 [0] NCCL INFO Using network Libfabric +ip-26-0-174-100:70020:70251 [1] NCCL INFO Using network Libfabric +ip-26-0-174-100:70019:70250 [0] NCCL INFO comm 0xbb0bce0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3e4adeaa2ec7858a - Init START +ip-26-0-174-100:70020:70251 [1] NCCL INFO comm 0x8d70cb0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e4adeaa2ec7858a - Init START +ip-26-0-174-100:70020:70251 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70019:70250 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/167 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-174-100:70020:70251 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-100:70020:70251 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-100:70020:70251 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-100:70019:70250 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70251 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70019:70250 [0] NCCL INFO Connected all rings +ip-26-0-174-100:70019:70250 [0] NCCL INFO Connected all trees +ip-26-0-174-100:70019:70250 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70019:70250 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70019:70250 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70020:70251 [1] NCCL INFO Connected all rings +ip-26-0-174-100:70020:70251 [1] NCCL INFO Connected all trees +ip-26-0-174-100:70020:70251 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70020:70251 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70020:70251 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70019:70250 [0] NCCL INFO comm 0xbb0bce0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3e4adeaa2ec7858a - Init COMPLETE +ip-26-0-174-100:70020:70251 [1] NCCL INFO comm 0x8d70cb0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e4adeaa2ec7858a - Init COMPLETE +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 00/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 01/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 04/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 05/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 06/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 07/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 08/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 09/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 10/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 11/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 12/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 13/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 14/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 15/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 16/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 17/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 18/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 19/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 20/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 21/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 22/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 23/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 24/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 25/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 26/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 27/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 28/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 29/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 30/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70257 [0] NCCL INFO Channel 31/1 : 0[0] -> 1[1] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Using network Libfabric +ip-26-0-174-100:70021:70261 [2] NCCL INFO Using network Libfabric +ip-26-0-174-100:70021:70261 [2] NCCL INFO comm 0x82df200 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x35fe81d9b0c564f6 - Init START +ip-26-0-174-100:70020:70260 [1] NCCL INFO comm 0xd06e860 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x35fe81d9b0c564f6 - Init START +ip-26-0-174-100:70020:70260 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70021:70261 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-174-100:70021:70261 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-100:70021:70261 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-100:70021:70261 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-100:70020:70260 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70261 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70260 [1] NCCL INFO Connected all rings +ip-26-0-174-100:70020:70260 [1] NCCL INFO Connected all trees +ip-26-0-174-100:70020:70260 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70020:70260 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70020:70260 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70021:70261 [2] NCCL INFO Connected all rings +ip-26-0-174-100:70021:70261 [2] NCCL INFO Connected all trees +ip-26-0-174-100:70021:70261 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70021:70261 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70021:70261 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70020:70260 [1] NCCL INFO comm 0xd06e860 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x35fe81d9b0c564f6 - Init COMPLETE +ip-26-0-174-100:70021:70261 [2] NCCL INFO comm 0x82df200 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x35fe81d9b0c564f6 - Init COMPLETE +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 00/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 01/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 04/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 05/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 06/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 07/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 08/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 09/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 10/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 11/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 12/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 13/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 14/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 15/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 16/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 17/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 18/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 19/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 20/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 21/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 22/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 23/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 24/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 25/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 26/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 27/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 28/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 29/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 30/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-100:70020:70268 [1] NCCL INFO Channel 31/1 : 0[1] -> 1[2] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Using network Libfabric +ip-26-0-174-100:70022:70273 [3] NCCL INFO Using network Libfabric +ip-26-0-174-100:70022:70273 [3] NCCL INFO comm 0x7ce7580 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb1bcb954f0df511e - Init START +ip-26-0-174-100:70021:70272 [2] NCCL INFO comm 0xc5dcea0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb1bcb954f0df511e - Init START +ip-26-0-174-100:70022:70273 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70021:70272 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70022:70273 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-174-100:70021:70272 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-100:70022:70273 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-100:70021:70272 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-100:70022:70273 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70021:70272 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70272 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70022:70273 [3] NCCL INFO Connected all rings +ip-26-0-174-100:70022:70273 [3] NCCL INFO Connected all trees +ip-26-0-174-100:70022:70273 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70022:70273 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70022:70273 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70021:70272 [2] NCCL INFO Connected all rings +ip-26-0-174-100:70021:70272 [2] NCCL INFO Connected all trees +ip-26-0-174-100:70021:70272 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70021:70272 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70021:70272 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70021:70272 [2] NCCL INFO comm 0xc5dcea0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb1bcb954f0df511e - Init COMPLETE +ip-26-0-174-100:70022:70273 [3] NCCL INFO comm 0x7ce7580 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb1bcb954f0df511e - Init COMPLETE +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 00/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 01/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 04/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 05/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 06/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 07/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 08/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 09/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 10/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 11/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 12/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 13/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 14/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 15/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 16/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 17/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 18/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 19/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 20/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 21/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 22/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 23/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 24/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 25/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 26/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 27/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 28/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 29/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 30/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-100:70021:70278 [2] NCCL INFO Channel 31/1 : 0[2] -> 1[3] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Using network Libfabric +ip-26-0-174-100:70023:70282 [4] NCCL INFO Using network Libfabric +ip-26-0-174-100:70023:70282 [4] NCCL INFO comm 0x7fd0e50 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x70a4ba3ee20a69f - Init START +ip-26-0-174-100:70022:70281 [3] NCCL INFO comm 0xbfe5b20 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x70a4ba3ee20a69f - Init START +ip-26-0-174-100:70023:70282 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70022:70281 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70023:70282 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-100:70023:70282 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-100:70023:70282 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-100:70022:70281 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 04/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 04/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 05/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 05/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 06/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 06/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 07/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 07/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 08/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 08/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 09/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 09/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 10/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 10/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 11/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 11/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 12/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 12/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 13/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 13/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 14/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 14/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 15/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 15/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 16/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 16/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 17/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 17/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 18/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 18/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 19/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 19/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 20/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 20/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 21/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 21/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 22/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 22/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Channel 23/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70281 [3] NCCL INFO Channel 23/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70023:70282 [4] NCCL INFO Connected all rings +ip-26-0-174-100:70023:70282 [4] NCCL INFO Connected all trees +ip-26-0-174-100:70023:70282 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70023:70282 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70023:70282 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70022:70281 [3] NCCL INFO Connected all rings +ip-26-0-174-100:70022:70281 [3] NCCL INFO Connected all trees +ip-26-0-174-100:70022:70281 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70022:70281 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70022:70281 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70023:70282 [4] NCCL INFO comm 0x7fd0e50 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x70a4ba3ee20a69f - Init COMPLETE +ip-26-0-174-100:70022:70281 [3] NCCL INFO comm 0xbfe5b20 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x70a4ba3ee20a69f - Init COMPLETE +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 00/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 01/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 04/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 05/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 06/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 07/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 08/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 09/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 10/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 11/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 12/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 13/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 14/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 15/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 16/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 17/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 18/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 19/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 20/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 21/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 22/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 23/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 24/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 25/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 26/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 27/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 28/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 29/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 30/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-100:70022:70289 [3] NCCL INFO Channel 31/1 : 0[3] -> 1[4] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Using network Libfabric +ip-26-0-174-100:70024:70292 [5] NCCL INFO Using network Libfabric +ip-26-0-174-100:70023:70291 [4] NCCL INFO comm 0xc2cebe0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x48d2d2ab0a68be0f - Init START +ip-26-0-174-100:70024:70292 [5] NCCL INFO comm 0x87f7e70 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x48d2d2ab0a68be0f - Init START +ip-26-0-174-100:70024:70292 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70023:70291 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70024:70292 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-100:70024:70292 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-100:70024:70292 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-100:70023:70291 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70291 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70024:70292 [5] NCCL INFO Connected all rings +ip-26-0-174-100:70024:70292 [5] NCCL INFO Connected all trees +ip-26-0-174-100:70024:70292 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70024:70292 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70024:70292 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70023:70291 [4] NCCL INFO Connected all rings +ip-26-0-174-100:70023:70291 [4] NCCL INFO Connected all trees +ip-26-0-174-100:70023:70291 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70023:70291 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70023:70291 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70024:70292 [5] NCCL INFO comm 0x87f7e70 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x48d2d2ab0a68be0f - Init COMPLETE +ip-26-0-174-100:70023:70291 [4] NCCL INFO comm 0xc2cebe0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x48d2d2ab0a68be0f - Init COMPLETE +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 00/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 01/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 04/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 05/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 06/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 07/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 08/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 09/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 10/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 11/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 12/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 13/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 14/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 15/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 16/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 17/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 18/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 19/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 20/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 21/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 22/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 23/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 24/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 25/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 26/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 27/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 28/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 29/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 30/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-100:70023:70299 [4] NCCL INFO Channel 31/1 : 0[4] -> 1[5] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Using network Libfabric +ip-26-0-174-100:70025:70302 [6] NCCL INFO Using network Libfabric +ip-26-0-174-100:70024:70301 [5] NCCL INFO comm 0xcaf5e60 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x654b348b6638649a - Init START +ip-26-0-174-100:70025:70302 [6] NCCL INFO comm 0x93c68a0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x654b348b6638649a - Init START +ip-26-0-174-100:70025:70302 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70024:70301 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70025:70302 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-100:70025:70302 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-100:70025:70302 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-100:70024:70301 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 04/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 05/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 04/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 06/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 05/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 07/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 06/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 08/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 07/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 09/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 08/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 10/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 09/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 11/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 10/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 12/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 11/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 13/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 12/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 14/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 13/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 15/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 14/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 16/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 15/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 17/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 16/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 18/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 17/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 19/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 18/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 20/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 19/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 21/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 20/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 22/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 21/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Channel 23/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 22/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70301 [5] NCCL INFO Channel 23/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70025:70302 [6] NCCL INFO Connected all rings +ip-26-0-174-100:70025:70302 [6] NCCL INFO Connected all trees +ip-26-0-174-100:70025:70302 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70025:70302 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70025:70302 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70024:70301 [5] NCCL INFO Connected all rings +ip-26-0-174-100:70024:70301 [5] NCCL INFO Connected all trees +ip-26-0-174-100:70024:70301 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70024:70301 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70024:70301 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70025:70302 [6] NCCL INFO comm 0x93c68a0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x654b348b6638649a - Init COMPLETE +ip-26-0-174-100:70024:70301 [5] NCCL INFO comm 0xcaf5e60 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x654b348b6638649a - Init COMPLETE +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 00/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 01/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 04/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 05/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 06/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 07/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 08/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 09/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 10/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 11/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 12/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 13/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 14/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 15/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 16/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 17/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 18/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 19/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 20/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 21/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 22/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 23/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 24/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 25/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 26/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 27/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 28/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 29/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 30/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-100:70024:70308 [5] NCCL INFO Channel 31/1 : 0[5] -> 1[6] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Using network Libfabric +ip-26-0-174-100:70026:70311 [7] NCCL INFO Using network Libfabric +ip-26-0-174-100:70025:70310 [6] NCCL INFO comm 0xd6c4820 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe22825062ae7c7af - Init START +ip-26-0-174-100:70026:70311 [7] NCCL INFO comm 0x906d020 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe22825062ae7c7af - Init START +ip-26-0-174-100:70025:70310 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70026:70311 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70026:70311 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-100:70026:70311 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-100:70026:70311 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-100:70025:70310 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70311 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70310 [6] NCCL INFO Connected all rings +ip-26-0-174-100:70025:70310 [6] NCCL INFO Connected all trees +ip-26-0-174-100:70025:70310 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70025:70310 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70025:70310 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70026:70311 [7] NCCL INFO Connected all rings +ip-26-0-174-100:70026:70311 [7] NCCL INFO Connected all trees +ip-26-0-174-100:70026:70311 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70026:70311 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70026:70311 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-100:70025:70310 [6] NCCL INFO comm 0xd6c4820 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe22825062ae7c7af - Init COMPLETE +ip-26-0-174-100:70026:70311 [7] NCCL INFO comm 0x906d020 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe22825062ae7c7af - Init COMPLETE +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 00/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 01/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 04/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 05/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 06/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 07/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 08/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 09/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 10/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 11/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 12/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 13/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 14/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 15/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 16/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 17/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 18/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 19/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 20/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 21/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 22/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 23/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 24/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 25/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 26/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 27/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 28/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 29/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 30/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70025:70318 [6] NCCL INFO Channel 31/1 : 0[6] -> 1[7] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-100:70026:70330 [7] NCCL INFO Using network Libfabric +ip-26-0-174-186:375438:375674 [0] NCCL INFO Using network Libfabric +ip-26-0-174-186:375438:375674 [0] NCCL INFO comm 0x7a1d780 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xeda1ff3c1c2ffaae - Init START +ip-26-0-174-100:70026:70330 [7] NCCL INFO comm 0xd36aaa0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xeda1ff3c1c2ffaae - Init START +ip-26-0-174-100:70026:70330 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375438:375674 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375438:375674 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-174-100:70026:70330 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70026:70330 [7] NCCL INFO Channel 00/02 : 0 1 +ip-26-0-174-100:70026:70330 [7] NCCL INFO Channel 01/02 : 0 1 +ip-26-0-174-100:70026:70330 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] -1/-1/-1->0->1 +ip-26-0-174-100:70026:70330 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375438:375674 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] 0/-1/-1->1->-1 +ip-26-0-174-186:375438:375674 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375438:375674 [0] NCCL INFO Channel 00/0 : 0[7] -> 1[0] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375438:375674 [0] NCCL INFO Channel 01/0 : 0[7] -> 1[0] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375438:375674 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[7] [send] via NET/Libfabric/1 +ip-26-0-174-100:70026:70330 [7] NCCL INFO Channel 00/0 : 1[0] -> 0[7] [receive] via NET/Libfabric/4 +ip-26-0-174-186:375438:375674 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[7] [send] via NET/Libfabric/1 +ip-26-0-174-100:70026:70330 [7] NCCL INFO Channel 01/0 : 1[0] -> 0[7] [receive] via NET/Libfabric/4 +ip-26-0-174-100:70026:70330 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[0] [send] via NET/Libfabric/4 +ip-26-0-174-100:70026:70330 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[0] [send] via NET/Libfabric/4 +ip-26-0-174-100:70026:70330 [7] NCCL INFO Connected all rings +ip-26-0-174-100:70026:70330 [7] NCCL INFO Connected all trees +ip-26-0-174-100:70026:70330 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375438:375674 [0] NCCL INFO Connected all rings +ip-26-0-174-186:375438:375674 [0] NCCL INFO Connected all trees +ip-26-0-174-100:70026:70330 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-100:70026:70330 [7] NCCL INFO 2 coll channels, 0 nvls channels, 2 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375438:375674 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375438:375674 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375438:375674 [0] NCCL INFO 2 coll channels, 0 nvls channels, 2 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375438:375674 [0] NCCL INFO comm 0x7a1d780 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xeda1ff3c1c2ffaae - Init COMPLETE +ip-26-0-174-100:70026:70330 [7] NCCL INFO comm 0xd36aaa0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xeda1ff3c1c2ffaae - Init COMPLETE +ip-26-0-174-100:70026:70334 [7] NCCL INFO Channel 00/1 : 0[7] -> 1[0] [send] via NET/Libfabric/7/Shared +ip-26-0-174-100:70026:70334 [7] NCCL INFO Channel 01/1 : 0[7] -> 1[0] [send] via NET/Libfabric/4/Shared +ip-26-0-174-186:375438:375678 [0] NCCL INFO Channel 00/1 : 0[7] -> 1[0] [receive] via NET/Libfabric/0/Shared +ip-26-0-174-186:375438:375678 [0] NCCL INFO Channel 01/1 : 0[7] -> 1[0] [receive] via NET/Libfabric/1/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Using network Libfabric +ip-26-0-174-186:375439:375681 [1] NCCL INFO Using network Libfabric +ip-26-0-174-186:375439:375681 [1] NCCL INFO comm 0x77c8c40 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc6c7b4ee59b4f907 - Init START +ip-26-0-174-186:375438:375680 [0] NCCL INFO comm 0xbd1b320 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc6c7b4ee59b4f907 - Init START +ip-26-0-174-186:375438:375680 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375439:375681 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-174-186:375439:375681 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-186:375439:375681 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-186:375439:375681 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-186:375438:375680 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375681 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375438:375680 [0] NCCL INFO Connected all rings +ip-26-0-174-186:375438:375680 [0] NCCL INFO Connected all trees +ip-26-0-174-186:375438:375680 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375438:375680 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375438:375680 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375439:375681 [1] NCCL INFO Connected all rings +ip-26-0-174-186:375439:375681 [1] NCCL INFO Connected all trees +ip-26-0-174-186:375439:375681 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375439:375681 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375439:375681 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375438:375680 [0] NCCL INFO comm 0xbd1b320 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc6c7b4ee59b4f907 - Init COMPLETE +ip-26-0-174-186:375439:375681 [1] NCCL INFO comm 0x77c8c40 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc6c7b4ee59b4f907 - Init COMPLETE +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 00/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 01/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 04/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 05/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 06/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 07/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 08/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 09/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 10/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 11/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 12/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 13/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 14/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 15/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 16/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 17/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 18/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 19/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 20/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 21/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 22/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 23/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 24/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 25/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 26/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 27/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 28/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 29/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 30/1 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-186:375438:375687 [0] NCCL INFO Channel 31/1 : 0[0] -> 1[1] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Using network Libfabric +ip-26-0-174-186:375440:375691 [2] NCCL INFO Using network Libfabric +ip-26-0-174-186:375440:375691 [2] NCCL INFO comm 0x81dab90 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8ec2a1cb44e7618a - Init START +ip-26-0-174-186:375439:375690 [1] NCCL INFO comm 0xbac6160 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8ec2a1cb44e7618a - Init START +ip-26-0-174-186:375439:375690 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375440:375691 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375440:375691 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-174-186:375439:375690 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-186:375440:375691 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-186:375440:375691 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-186:375439:375690 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375690 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375440:375691 [2] NCCL INFO Connected all rings +ip-26-0-174-186:375440:375691 [2] NCCL INFO Connected all trees +ip-26-0-174-186:375440:375691 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375440:375691 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375440:375691 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375439:375690 [1] NCCL INFO Connected all rings +ip-26-0-174-186:375439:375690 [1] NCCL INFO Connected all trees +ip-26-0-174-186:375439:375690 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375439:375690 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375439:375690 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375440:375691 [2] NCCL INFO comm 0x81dab90 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8ec2a1cb44e7618a - Init COMPLETE +ip-26-0-174-186:375439:375690 [1] NCCL INFO comm 0xbac6160 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8ec2a1cb44e7618a - Init COMPLETE +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 00/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 01/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 04/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 05/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 06/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 07/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 08/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 09/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 10/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 11/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 12/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 13/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 14/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 15/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 16/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 17/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 18/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 19/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 20/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 21/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 22/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 23/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 24/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 25/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 26/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 27/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 28/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 29/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 30/1 : 0[1] -> 1[2] via P2P/IPC +ip-26-0-174-186:375439:375696 [1] NCCL INFO Channel 31/1 : 0[1] -> 1[2] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Using network Libfabric +ip-26-0-174-186:375441:375701 [3] NCCL INFO Using network Libfabric +ip-26-0-174-186:375441:375701 [3] NCCL INFO comm 0x7d82dc0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x412ea6009b8a339 - Init START +ip-26-0-174-186:375440:375700 [2] NCCL INFO comm 0xc4d8630 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x412ea6009b8a339 - Init START +ip-26-0-174-186:375440:375700 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375441:375701 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375441:375701 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-174-186:375440:375700 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-186:375441:375701 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-186:375440:375700 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375441:375701 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375700 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375441:375701 [3] NCCL INFO Connected all rings +ip-26-0-174-186:375441:375701 [3] NCCL INFO Connected all trees +ip-26-0-174-186:375441:375701 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375441:375701 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375441:375701 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375440:375700 [2] NCCL INFO Connected all rings +ip-26-0-174-186:375440:375700 [2] NCCL INFO Connected all trees +ip-26-0-174-186:375440:375700 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375440:375700 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375440:375700 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375441:375701 [3] NCCL INFO comm 0x7d82dc0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x412ea6009b8a339 - Init COMPLETE +ip-26-0-174-186:375440:375700 [2] NCCL INFO comm 0xc4d8630 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x412ea6009b8a339 - Init COMPLETE +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 00/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 01/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 04/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 05/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 06/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 07/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 08/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 09/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 10/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 11/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 12/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 13/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 14/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 15/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 16/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 17/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 18/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 19/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 20/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 21/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 22/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 23/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 24/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 25/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 26/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 27/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 28/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 29/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 30/1 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-174-186:375440:375707 [2] NCCL INFO Channel 31/1 : 0[2] -> 1[3] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Using network Libfabric +ip-26-0-174-186:375442:375711 [4] NCCL INFO Using network Libfabric +ip-26-0-174-186:375442:375711 [4] NCCL INFO comm 0x7fd40f0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xbf8ca8d0bcf3de9e - Init START +ip-26-0-174-186:375441:375710 [3] NCCL INFO comm 0xc080ed0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbf8ca8d0bcf3de9e - Init START +ip-26-0-174-186:375442:375711 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375441:375710 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-174-186:375442:375711 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375442:375711 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-186:375442:375711 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-186:375441:375710 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 04/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 05/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 04/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 06/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 05/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 07/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 06/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 08/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 07/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 09/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 08/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 10/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 09/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 11/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 10/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 12/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 11/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 13/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 12/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 14/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 13/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 15/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 14/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 16/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 15/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 17/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 16/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 18/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 17/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 19/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 18/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 20/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 19/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 21/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 20/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 22/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 21/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Channel 23/0 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 22/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375710 [3] NCCL INFO Channel 23/0 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375442:375711 [4] NCCL INFO Connected all rings +ip-26-0-174-186:375442:375711 [4] NCCL INFO Connected all trees +ip-26-0-174-186:375442:375711 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375442:375711 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375442:375711 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375441:375710 [3] NCCL INFO Connected all rings +ip-26-0-174-186:375441:375710 [3] NCCL INFO Connected all trees +ip-26-0-174-186:375441:375710 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375441:375710 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375441:375710 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375442:375711 [4] NCCL INFO comm 0x7fd40f0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xbf8ca8d0bcf3de9e - Init COMPLETE +ip-26-0-174-186:375441:375710 [3] NCCL INFO comm 0xc080ed0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbf8ca8d0bcf3de9e - Init COMPLETE +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 00/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 01/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 04/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 05/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 06/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 07/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 08/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 09/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 10/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 11/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 12/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 13/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 14/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 15/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 16/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 17/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 18/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 19/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 20/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 21/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 22/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 23/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 24/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 25/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 26/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 27/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 28/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 29/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 30/1 : 0[3] -> 1[4] via P2P/IPC +ip-26-0-174-186:375441:375719 [3] NCCL INFO Channel 31/1 : 0[3] -> 1[4] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Using network Libfabric +ip-26-0-174-186:375443:375722 [5] NCCL INFO Using network Libfabric +ip-26-0-174-186:375443:375722 [5] NCCL INFO comm 0x75d65c0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xab9d025686870fe2 - Init START +ip-26-0-174-186:375442:375721 [4] NCCL INFO comm 0xc2d2b90 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xab9d025686870fe2 - Init START +ip-26-0-174-186:375442:375721 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375443:375722 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375443:375722 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-186:375443:375722 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-186:375443:375722 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-186:375442:375721 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375721 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375443:375722 [5] NCCL INFO Connected all rings +ip-26-0-174-186:375443:375722 [5] NCCL INFO Connected all trees +ip-26-0-174-186:375443:375722 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375443:375722 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375443:375722 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375442:375721 [4] NCCL INFO Connected all rings +ip-26-0-174-186:375442:375721 [4] NCCL INFO Connected all trees +ip-26-0-174-186:375442:375721 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375442:375721 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375442:375721 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375443:375722 [5] NCCL INFO comm 0x75d65c0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xab9d025686870fe2 - Init COMPLETE +ip-26-0-174-186:375442:375721 [4] NCCL INFO comm 0xc2d2b90 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xab9d025686870fe2 - Init COMPLETE +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 00/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 01/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 04/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 05/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 06/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 07/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 08/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 09/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 10/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 11/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 12/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 13/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 14/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 15/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 16/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 17/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 18/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 19/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 20/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 21/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 22/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 23/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 24/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 25/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 26/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 27/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 28/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 29/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 30/1 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-174-186:375442:375727 [4] NCCL INFO Channel 31/1 : 0[4] -> 1[5] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Using network Libfabric +ip-26-0-174-186:375444:375731 [6] NCCL INFO Using network Libfabric +ip-26-0-174-186:375443:375730 [5] NCCL INFO comm 0xb8d45d0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4848aded38b7b116 - Init START +ip-26-0-174-186:375444:375731 [6] NCCL INFO comm 0x874d430 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4848aded38b7b116 - Init START +ip-26-0-174-186:375444:375731 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375443:375730 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375444:375731 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-186:375443:375730 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375444:375731 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-186:375444:375731 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 04/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 05/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 04/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 06/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 05/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 07/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 06/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 08/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 07/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 09/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 08/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 10/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 09/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 11/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 10/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 12/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 11/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 13/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 12/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 14/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 13/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 15/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 14/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 16/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 15/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 17/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 16/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 18/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 17/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 19/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 18/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 20/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 19/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 21/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 20/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 22/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 21/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Channel 23/0 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 22/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375731 [6] NCCL INFO Channel 23/0 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375730 [5] NCCL INFO Connected all rings +ip-26-0-174-186:375443:375730 [5] NCCL INFO Connected all trees +ip-26-0-174-186:375443:375730 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375443:375730 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375443:375730 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375444:375731 [6] NCCL INFO Connected all rings +ip-26-0-174-186:375444:375731 [6] NCCL INFO Connected all trees +ip-26-0-174-186:375444:375731 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375444:375731 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375444:375731 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375443:375730 [5] NCCL INFO comm 0xb8d45d0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4848aded38b7b116 - Init COMPLETE +ip-26-0-174-186:375444:375731 [6] NCCL INFO comm 0x874d430 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4848aded38b7b116 - Init COMPLETE +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 00/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 01/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 04/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 05/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 06/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 07/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 08/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 09/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 10/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 11/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 12/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 13/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 14/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 15/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 16/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 17/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 18/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 19/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 20/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 21/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 22/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 23/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 24/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 25/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 26/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 27/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 28/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 29/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 30/1 : 0[5] -> 1[6] via P2P/IPC +ip-26-0-174-186:375443:375737 [5] NCCL INFO Channel 31/1 : 0[5] -> 1[6] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Using network Libfabric +ip-26-0-174-186:375445:375741 [7] NCCL INFO Using network Libfabric +ip-26-0-174-186:375445:375741 [7] NCCL INFO comm 0x8f1c990 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x20d293ac5c2341c4 - Init START +ip-26-0-174-186:375444:375740 [6] NCCL INFO comm 0xca4b1b0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x20d293ac5c2341c4 - Init START +ip-26-0-174-186:375444:375740 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375445:375741 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375445:375741 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-174-186:375445:375741 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-174-186:375445:375741 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-174-186:375444:375740 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375741 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375740 [6] NCCL INFO Connected all rings +ip-26-0-174-186:375444:375740 [6] NCCL INFO Connected all trees +ip-26-0-174-186:375444:375740 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375444:375740 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375444:375740 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375445:375741 [7] NCCL INFO Connected all rings +ip-26-0-174-186:375445:375741 [7] NCCL INFO Connected all trees +ip-26-0-174-186:375445:375741 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375445:375741 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-186:375445:375741 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-174-186:375444:375740 [6] NCCL INFO comm 0xca4b1b0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x20d293ac5c2341c4 - Init COMPLETE +ip-26-0-174-186:375445:375741 [7] NCCL INFO comm 0x8f1c990 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x20d293ac5c2341c4 - Init COMPLETE +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 00/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 01/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 04/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 05/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 06/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 07/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 08/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 09/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 10/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 11/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 12/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 13/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 14/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 15/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 16/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 17/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 18/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 19/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 20/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 21/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 22/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 23/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 24/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 25/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 26/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 27/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 28/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 29/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 30/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-186:375444:375747 [6] NCCL INFO Channel 31/1 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-174-100:70026:70404 [7] NCCL INFO Channel 00/1 : 1[0] -> 0[7] [receive] via NET/Libfabric/7/Shared +ip-26-0-174-100:70026:70404 [7] NCCL INFO Channel 01/1 : 1[0] -> 0[7] [receive] via NET/Libfabric/4/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 00/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 01/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 04/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 05/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 06/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 07/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 08/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 09/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 10/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 11/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 12/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 13/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 14/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 15/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 16/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 17/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 18/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 19/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 20/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 21/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 22/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 23/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 24/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 25/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 26/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 27/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 28/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 29/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 30/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375445:375820 [7] NCCL INFO Channel 31/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 00/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 01/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 04/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 05/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 06/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 07/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 08/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 09/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 10/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 11/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 12/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 13/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 14/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 15/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 16/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 17/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 18/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 19/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 20/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 21/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 22/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 23/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 24/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 25/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 26/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 27/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 28/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 29/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 30/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375444:375822 [6] NCCL INFO Channel 31/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 00/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 01/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 04/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 05/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 06/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 07/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 08/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 09/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 10/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 11/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 12/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 13/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 14/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 15/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 16/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 17/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 18/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 19/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 20/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 21/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 22/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 23/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 24/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 25/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 26/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 27/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 28/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 29/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 30/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375443:375823 [5] NCCL INFO Channel 31/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 00/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 01/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 04/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 05/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 06/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 07/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 08/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 09/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 10/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 11/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 12/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 13/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 14/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 15/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 16/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 17/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 18/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 19/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 20/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 21/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 22/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 23/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 24/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 25/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 26/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 27/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 28/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 29/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 30/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375442:375824 [4] NCCL INFO Channel 31/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 00/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 01/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 04/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 05/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 06/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 07/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 08/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 09/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 10/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 11/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 12/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 13/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 14/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 15/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 16/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 17/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 18/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 19/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 20/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 21/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 22/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 23/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 24/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 25/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 26/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 27/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 28/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 29/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 30/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375441:375825 [3] NCCL INFO Channel 31/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 00/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 01/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 04/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 05/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 06/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 07/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 08/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 09/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 10/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 11/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 12/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 13/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 14/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 15/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 16/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 17/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 18/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 19/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 20/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 21/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 22/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 23/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 24/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 25/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 26/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 27/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 28/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 29/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 30/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375440:375827 [2] NCCL INFO Channel 31/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 00/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 01/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 04/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 05/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 06/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 07/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 08/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 09/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 10/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 11/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 12/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 13/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 14/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 15/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 16/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 17/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 18/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 19/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 20/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 21/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 22/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 23/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 24/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 25/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 26/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 27/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 28/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 29/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 30/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375439:375828 [1] NCCL INFO Channel 31/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375438:375830 [0] NCCL INFO Channel 00/1 : 1[0] -> 0[7] [send] via NET/Libfabric/0/Shared +ip-26-0-174-186:375438:375830 [0] NCCL INFO Channel 01/1 : 1[0] -> 0[7] [send] via NET/Libfabric/1/Shared +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 00/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 01/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 04/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 05/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 06/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 07/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 08/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 09/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 10/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 11/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 12/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 13/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 14/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 15/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 16/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 17/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 18/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 19/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 20/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 21/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 22/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 23/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 24/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 25/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 26/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 27/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 28/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 29/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 30/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70026:70407 [7] NCCL INFO Channel 31/1 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 00/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 01/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 04/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 05/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 06/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 07/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 08/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 09/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 10/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 11/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 12/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 13/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 14/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 15/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 16/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 17/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 18/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 19/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 20/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 21/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 22/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 23/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 24/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 25/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 26/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 27/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 28/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 29/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 30/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70025:70408 [6] NCCL INFO Channel 31/1 : 1[6] -> 0[5] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 00/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 01/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 04/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 05/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 06/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 07/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 08/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 09/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 10/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 11/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 12/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 13/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 14/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 15/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 16/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 17/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 18/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 19/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 20/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 21/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 22/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 23/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 24/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 25/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 26/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 27/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 28/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 29/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 30/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70024:70410 [5] NCCL INFO Channel 31/1 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 00/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 01/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 04/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 05/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 06/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 07/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 08/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 09/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 10/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 11/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 12/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 13/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 14/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 15/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 16/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 17/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 18/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 19/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 20/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 21/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 22/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 23/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 24/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 25/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 26/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 27/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 28/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 29/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 30/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70023:70411 [4] NCCL INFO Channel 31/1 : 1[4] -> 0[3] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 00/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 01/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 04/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 05/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 06/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 07/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 08/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 09/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 10/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 11/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 12/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 13/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 14/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 15/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 16/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 17/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 18/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 19/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 20/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 21/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 22/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 23/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 24/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 25/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 26/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 27/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 28/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 29/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 30/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70022:70413 [3] NCCL INFO Channel 31/1 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 00/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 01/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 04/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 05/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 06/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 07/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 08/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 09/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 10/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 11/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 12/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 13/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 14/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 15/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 16/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 17/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 18/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 19/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 20/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 21/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 22/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 23/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 24/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 25/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 26/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 27/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 28/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 29/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 30/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70021:70414 [2] NCCL INFO Channel 31/1 : 1[2] -> 0[1] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 00/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 01/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 04/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 05/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 06/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 07/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 08/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 09/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 10/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 11/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 12/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 13/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 14/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 15/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 16/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 17/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 18/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 19/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 20/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 21/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 22/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 23/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 24/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 25/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 26/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 27/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 28/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 29/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 30/1 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70415 [1] NCCL INFO Channel 31/1 : 1[1] -> 0[0] via P2P/IPC +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +ip-26-0-174-100:70019:70417 [0] NCCL INFO Using network Libfabric +ip-26-0-174-186:375439:375837 [1] NCCL INFO Using network Libfabric +ip-26-0-174-186:375445:375832 [7] NCCL INFO Using network Libfabric +ip-26-0-174-186:375443:375836 [5] NCCL INFO Using network Libfabric +ip-26-0-174-186:375442:375838 [4] NCCL INFO Using network Libfabric +ip-26-0-174-100:70025:70419 [6] NCCL INFO Using network Libfabric +ip-26-0-174-186:375438:375833 [0] NCCL INFO Using network Libfabric +ip-26-0-174-186:375440:375835 [2] NCCL INFO Using network Libfabric +ip-26-0-174-100:70021:70422 [2] NCCL INFO Using network Libfabric +ip-26-0-174-186:375441:375839 [3] NCCL INFO Using network Libfabric +ip-26-0-174-100:70024:70421 [5] NCCL INFO Using network Libfabric +ip-26-0-174-100:70023:70418 [4] NCCL INFO Using network Libfabric +ip-26-0-174-186:375444:375834 [6] NCCL INFO Using network Libfabric +ip-26-0-174-100:70020:70420 [1] NCCL INFO Using network Libfabric +ip-26-0-174-100:70026:70423 [7] NCCL INFO Using network Libfabric +ip-26-0-174-100:70022:70424 [3] NCCL INFO Using network Libfabric +ip-26-0-174-100:70022:70424 [3] NCCL INFO comm 0xcfb0380 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-100:70026:70423 [7] NCCL INFO comm 0xe334c20 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-100:70025:70419 [6] NCCL INFO comm 0xe68f630 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-100:70021:70422 [2] NCCL INFO comm 0xd5a7e20 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-100:70024:70421 [5] NCCL INFO comm 0xdac0e20 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-100:70023:70418 [4] NCCL INFO comm 0xd299a00 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-186:375443:375836 [5] NCCL INFO comm 0xc89f1d0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-186:375445:375832 [7] NCCL INFO comm 0x100cc140 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-186:375442:375838 [4] NCCL INFO comm 0xd29e280 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-100:70020:70420 [1] NCCL INFO comm 0xe0398d0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-100:70019:70417 [0] NCCL INFO comm 0xe70c7c0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-186:375444:375834 [6] NCCL INFO comm 0xda161f0 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-186:375441:375839 [3] NCCL INFO comm 0xd04bab0 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-186:375440:375835 [2] NCCL INFO comm 0xd4a3930 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-186:375438:375833 [0] NCCL INFO comm 0xcce58f0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-186:375439:375837 [1] NCCL INFO comm 0xca91250 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd3aab9699a254c5 - Init START +ip-26-0-174-100:70022:70424 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70021:70422 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70026:70423 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70025:70419 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70020:70420 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70024:70421 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375445:375832 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375444:375834 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70023:70418 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70019:70417 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/167 +ip-26-0-174-186:375443:375836 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375442:375838 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375440:375835 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375441:375839 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375439:375837 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-186:375438:375833 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/146 +ip-26-0-174-100:70024:70421 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70024:70421 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-174-100:70025:70419 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70025:70419 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-174-186:375443:375836 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375443:375836 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-174-100:70021:70422 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-174-186:375441:375839 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-174-186:375445:375832 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375445:375832 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-174-186:375444:375834 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375444:375834 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-174-100:70023:70418 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70023:70418 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-174-186:375439:375837 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-174-100:70019:70417 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-174-186:375440:375835 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-174-186:375438:375833 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-174-100:70022:70424 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-174-100:70020:70420 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-174-100:70026:70423 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-174-100:70026:70423 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-174-186:375442:375838 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-174-186:375442:375838 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-174-100:70020:70420 [1] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70020:70420 [1] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70020:70420 [1] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70020:70420 [1] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70020:70420 [1] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Trees [0] 2/9/-1->1->-1 [1] 3/-1/-1->1->2 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->9 [5] 3/-1/-1->1->2 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375445:375832 [7] NCCL INFO Trees [0] 8/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->13 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->13 +ip-26-0-174-186:375445:375832 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70021:70422 [2] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70022:70424 [3] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70021:70422 [2] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70022:70424 [3] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70021:70422 [2] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70021:70422 [2] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70022:70424 [3] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70021:70422 [2] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70021:70422 [2] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70021:70422 [2] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70021:70422 [2] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 1/10/-1->2->-1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 1/-1/-1->2->10 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 +ip-26-0-174-100:70019:70417 [0] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->1 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->1 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 +ip-26-0-174-100:70021:70422 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70022:70424 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70019:70417 [0] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70019:70417 [0] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70023:70418 [4] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70019:70417 [0] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70023:70418 [4] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70019:70417 [0] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-186:375443:375836 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->5 [3] 15/-1/-1->13->14 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->12 [6] 14/5/-1->13->-1 [7] 15/-1/-1->13->14 +ip-26-0-174-186:375443:375836 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70023:70418 [4] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70026:70423 [7] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70019:70417 [0] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70023:70418 [4] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70026:70423 [7] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70019:70417 [0] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70026:70423 [7] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70023:70418 [4] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70019:70417 [0] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70026:70423 [7] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-186:375444:375834 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 13/-1/-1->14->6 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->13 [7] 13/6/-1->14->-1 +ip-26-0-174-186:375444:375834 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70023:70418 [4] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70026:70423 [7] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70025:70419 [6] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70023:70418 [4] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70026:70423 [7] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 00/08 : 0 7 6 5 4 3 2 1 8 15 14 13 12 11 10 9 +ip-26-0-174-100:70023:70418 [4] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70025:70419 [6] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70026:70423 [7] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 01/08 : 0 7 6 5 4 2 3 9 8 15 14 13 12 10 11 1 +ip-26-0-174-186:375442:375838 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] -1/-1/-1->12->11 [3] -1/-1/-1->12->11 [4] 13/-1/-1->12->11 [5] 13/-1/-1->12->11 [6] -1/-1/-1->12->11 [7] -1/-1/-1->12->11 +ip-26-0-174-186:375442:375838 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70026:70423 [7] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70025:70419 [6] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 02/08 : 0 2 7 6 5 12 11 9 8 10 15 14 13 4 3 1 +ip-26-0-174-100:70025:70419 [6] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70024:70421 [5] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-174-100:70025:70419 [6] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70023:70418 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] -1/-1/-1->4->3 [3] -1/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] -1/-1/-1->4->3 [7] -1/-1/-1->4->3 +ip-26-0-174-100:70024:70421 [5] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-174-100:70025:70419 [6] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 03/08 : 0 4 2 6 7 13 11 9 8 12 10 14 15 5 3 1 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->9 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->10 [4] 12/-1/-1->11->10 [5] 12/-1/-1->11->9 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Trees [0] -1/-1/-1->8->15 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] -1/-1/-1->8->15 [5] -1/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 9/-1/-1->10->2 [2] 11/-1/-1->10->9 [3] 11/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 9/2/-1->10->-1 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 +ip-26-0-174-186:375441:375839 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375438:375833 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375440:375835 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70026:70423 [7] NCCL INFO Trees [0] 0/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->5 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/-1/-1->7->5 +ip-26-0-174-100:70023:70418 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70024:70421 [5] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-174-100:70026:70423 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70025:70419 [6] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70024:70421 [5] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 04/08 : 0 7 6 5 4 3 2 1 8 15 14 13 12 11 10 9 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Trees [0] 10/-1/-1->9->1 [1] 11/-1/-1->9->10 [2] 10/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/1/-1->9->-1 [5] 11/-1/-1->9->10 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 +ip-26-0-174-186:375439:375837 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70025:70419 [6] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70024:70421 [5] NCCL INFO NVLS Head 4: 0 0 +ip-26-0-174-100:70024:70421 [5] NCCL INFO NVLS Head 5: 0 0 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 05/08 : 0 7 6 5 4 2 3 9 8 15 14 13 12 10 11 1 +ip-26-0-174-100:70024:70421 [5] NCCL INFO NVLS Head 6: 0 0 +ip-26-0-174-100:70024:70421 [5] NCCL INFO NVLS Head 7: 0 0 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 06/08 : 0 2 7 6 5 12 11 9 8 10 15 14 13 4 3 1 +ip-26-0-174-100:70025:70419 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 5/14/-1->6->-1 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 5/-1/-1->6->14 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 07/08 : 0 4 2 6 7 13 11 9 8 12 10 14 15 5 3 1 +ip-26-0-174-100:70025:70419 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70024:70421 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/13/-1->5->-1 [3] 7/-1/-1->5->6 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->13 [7] 7/-1/-1->5->6 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Trees [0] -1/-1/-1->0->7 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] -1/-1/-1->0->7 [5] -1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 +ip-26-0-174-100:70024:70421 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-100:70019:70417 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 02/0 : 8[0] -> 10[2] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 02/0 : 0[0] -> 2[2] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 06/0 : 8[0] -> 10[2] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 07/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 06/0 : 0[0] -> 2[2] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 03/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 03/0 : 10[2] -> 14[6] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 03/0 : 2[2] -> 6[6] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 03/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 01/0 : 11[3] -> 1[1] [send] via NET/Libfabric/3 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 05/0 : 11[3] -> 1[1] [send] via NET/Libfabric/3 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 07/0 : 2[2] -> 6[6] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 07/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 01/0 : 3[3] -> 9[1] [send] via NET/Libfabric/3 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 05/0 : 3[3] -> 9[1] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 07/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 07/0 : 10[2] -> 14[6] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 02/0 : 2[2] -> 7[7] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 06/0 : 2[2] -> 7[7] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 02/0 : 10[2] -> 15[7] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 06/0 : 10[2] -> 15[7] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 03/0 : 7[7] -> 13[5] [send] via NET/Libfabric/7 +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 07/0 : 7[7] -> 13[5] [send] via NET/Libfabric/7 +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 03/0 : 15[7] -> 5[5] [send] via NET/Libfabric/7 +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 07/0 : 15[7] -> 5[5] [send] via NET/Libfabric/7 +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 06/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 03/0 : 15[7] -> 5[5] [receive] via NET/Libfabric/6 +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 07/0 : 15[7] -> 5[5] [receive] via NET/Libfabric/6 +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 02/0 : 5[5] -> 12[4] [send] via NET/Libfabric/6 +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 06/0 : 5[5] -> 12[4] [send] via NET/Libfabric/6 +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 03/0 : 7[7] -> 13[5] [receive] via NET/Libfabric/6 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 01/0 : 11[3] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 07/0 : 7[7] -> 13[5] [receive] via NET/Libfabric/6 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 05/0 : 11[3] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 02/0 : 13[5] -> 4[4] [send] via NET/Libfabric/6 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 00/0 : 1[1] -> 8[0] [send] via NET/Libfabric/2 +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 06/0 : 13[5] -> 4[4] [send] via NET/Libfabric/6 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 04/0 : 1[1] -> 8[0] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 01/0 : 3[3] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 05/0 : 3[3] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 06/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 00/0 : 9[1] -> 0[0] [send] via NET/Libfabric/2 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 04/0 : 9[1] -> 0[0] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 07/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 00/0 : 9[1] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 04/0 : 9[1] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 06/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 00/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 07/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 00/0 : 1[1] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 04/0 : 1[1] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 00/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 02/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4 +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 06/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4 +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 02/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 06/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4 +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 07/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 03/0 : 4[4] -> 2[2] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 03/0 : 12[4] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 07/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 05/0 : 4[4] -> 2[2] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 05/0 : 12[4] -> 10[2] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 07/0 : 12[4] -> 10[2] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 07/0 : 4[4] -> 2[2] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 05/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Connected all rings +ip-26-0-174-186:375438:375833 [0] NCCL INFO Connected all rings +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Connected all rings +ip-26-0-174-100:70025:70419 [6] NCCL INFO Connected all rings +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Connected all rings +ip-26-0-174-100:70023:70418 [4] NCCL INFO Connected all rings +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Connected all rings +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 05/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Connected all rings +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Connected all rings +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 01/0 : 1[1] -> 3[3] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 05/0 : 1[1] -> 3[3] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Connected all rings +ip-26-0-174-186:375441:375839 [3] NCCL INFO Connected all rings +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Connected all rings +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Connected all rings +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 04/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 06/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 03/0 : 13[5] -> 15[7] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 07/0 : 13[5] -> 15[7] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Connected all rings +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-186:375442:375838 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 03/0 : 5[5] -> 7[7] via P2P/IPC +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 07/0 : 5[5] -> 7[7] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 03/0 : 7[7] -> 5[5] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Channel 07/0 : 7[7] -> 5[5] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Connected all rings +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 01/0 : 9[1] -> 11[3] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 05/0 : 9[1] -> 11[3] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 01/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 05/0 : 11[3] -> 9[1] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 00/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Connected all rings +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 03/0 : 15[7] -> 13[5] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Channel 07/0 : 15[7] -> 13[5] via P2P/IPC +ip-26-0-174-100:70019:70417 [0] NCCL INFO Connected all trees +ip-26-0-174-100:70019:70417 [0] NCCL INFO NVLS comm 0xe70c7c0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 01/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 05/0 : 3[3] -> 1[1] via P2P/IPC +ip-26-0-174-100:70023:70418 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-174-186:375438:375833 [0] NCCL INFO Connected all trees +ip-26-0-174-186:375438:375833 [0] NCCL INFO NVLS comm 0xcce58f0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/7 +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/7 +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/7 +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/7 +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/4 +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/4 +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [send] via NET/Libfabric/4 +ip-26-0-174-186:375443:375836 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [send] via NET/Libfabric/4 +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/4 +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/4 +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [send] via NET/Libfabric/4 +ip-26-0-174-100:70024:70421 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [send] via NET/Libfabric/4 +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/7 +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/7 +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/7 +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/7 +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-100:70025:70419 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-174-186:375444:375834 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/3 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [send] via NET/Libfabric/0 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/3 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/3 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-186:375445:375832 [7] NCCL INFO Connected all trees +ip-26-0-174-186:375445:375832 [7] NCCL INFO NVLS comm 0x100cc140 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-174-100:70026:70423 [7] NCCL INFO Connected all trees +ip-26-0-174-100:70026:70423 [7] NCCL INFO NVLS comm 0xe334c20 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70023:70418 [4] NCCL INFO Connected all trees +ip-26-0-174-100:70023:70418 [4] NCCL INFO NVLS comm 0xd299a00 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375442:375838 [4] NCCL INFO Connected all trees +ip-26-0-174-186:375442:375838 [4] NCCL INFO NVLS comm 0xd29e280 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70024:70421 [5] NCCL INFO Connected all trees +ip-26-0-174-100:70024:70421 [5] NCCL INFO NVLS comm 0xdac0e20 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70025:70419 [6] NCCL INFO Connected all trees +ip-26-0-174-100:70025:70419 [6] NCCL INFO NVLS comm 0xe68f630 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Connected all trees +ip-26-0-174-186:375441:375839 [3] NCCL INFO Connected all trees +ip-26-0-174-100:70022:70424 [3] NCCL INFO NVLS comm 0xcfb0380 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375444:375834 [6] NCCL INFO Connected all trees +ip-26-0-174-186:375443:375836 [5] NCCL INFO Connected all trees +ip-26-0-174-186:375441:375839 [3] NCCL INFO NVLS comm 0xd04bab0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375444:375834 [6] NCCL INFO NVLS comm 0xda161f0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375443:375836 [5] NCCL INFO NVLS comm 0xc89f1d0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Connected all trees +ip-26-0-174-100:70020:70420 [1] NCCL INFO NVLS comm 0xe0398d0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Connected all trees +ip-26-0-174-186:375439:375837 [1] NCCL INFO NVLS comm 0xca91250 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Connected all trees +ip-26-0-174-186:375440:375835 [2] NCCL INFO Connected all trees +ip-26-0-174-100:70021:70422 [2] NCCL INFO NVLS comm 0xd5a7e20 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-186:375440:375835 [2] NCCL INFO NVLS comm 0xd4a3930 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-174-100:70026:70423 [7] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70026:70423 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70026:70423 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70026:70423 [7] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70024:70421 [5] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70024:70421 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70024:70421 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70024:70421 [5] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70023:70418 [4] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70023:70418 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70023:70418 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70023:70418 [4] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70025:70419 [6] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70025:70419 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70025:70419 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70025:70419 [6] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375445:375832 [7] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375445:375832 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375445:375832 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375445:375832 [7] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375443:375836 [5] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375443:375836 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375443:375836 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375443:375836 [5] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375444:375834 [6] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375444:375834 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375444:375834 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375444:375834 [6] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375442:375838 [4] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375442:375838 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375442:375838 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375442:375838 [4] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70022:70424 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-100:70019:70417 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375439:375837 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/2 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375441:375839 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-186:375438:375833 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70021:70422 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-186:375440:375835 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/1 +ip-26-0-174-100:70020:70420 [1] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70020:70420 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375439:375837 [1] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70020:70420 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70020:70420 [1] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375439:375837 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375439:375837 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375439:375837 [1] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375441:375839 [3] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375441:375839 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375441:375839 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375441:375839 [3] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70022:70424 [3] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70022:70424 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70022:70424 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70022:70424 [3] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70019:70417 [0] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70019:70417 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70019:70417 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70019:70417 [0] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375438:375833 [0] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375438:375833 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375438:375833 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375438:375833 [0] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70021:70422 [2] NCCL INFO Connected NVLS tree +ip-26-0-174-100:70021:70422 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-100:70021:70422 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-100:70021:70422 [2] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-186:375440:375835 [2] NCCL INFO Connected NVLS tree +ip-26-0-174-186:375440:375835 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-174-186:375440:375835 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-174-186:375440:375835 [2] NCCL INFO 8 coll channels, 16 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-174-100:70022:70424 [3] NCCL INFO comm 0xcfb0380 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-100:70019:70417 [0] NCCL INFO comm 0xe70c7c0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-100:70026:70423 [7] NCCL INFO comm 0xe334c20 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-100:70020:70420 [1] NCCL INFO comm 0xe0398d0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-100:70025:70419 [6] NCCL INFO comm 0xe68f630 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-100:70021:70422 [2] NCCL INFO comm 0xd5a7e20 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-100:70023:70418 [4] NCCL INFO comm 0xd299a00 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-100:70024:70421 [5] NCCL INFO comm 0xdac0e20 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-186:375444:375834 [6] NCCL INFO comm 0xda161f0 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-186:375441:375839 [3] NCCL INFO comm 0xd04bab0 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-186:375438:375833 [0] NCCL INFO comm 0xcce58f0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-186:375440:375835 [2] NCCL INFO comm 0xd4a3930 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-186:375439:375837 [1] NCCL INFO comm 0xca91250 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-186:375443:375836 [5] NCCL INFO comm 0xc89f1d0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-186:375442:375838 [4] NCCL INFO comm 0xd29e280 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd3aab9699a254c5 - Init COMPLETE +ip-26-0-174-186:375445:375832 [7] NCCL INFO comm 0x100cc140 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd3aab9699a254c5 - Init COMPLETE +Successfully completed training iteration with AllForwardAllBackwardPipelineEngine +Successfully completed training iteration with OneForwardOneBackwardPipelineEngine +ip-26-0-174-100:70019:70212 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-174-100:70024:70209 [5] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-174-100:70019:70212 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-174-100:70019:70212 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70019:70212 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-100:70019:70212 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-174-100:70023:70206 [4] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-174-100:70020:70211 [1] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70026:70207 [7] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-174-100:70019:70212 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-174-100:70025:70208 [6] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-174-100:70021:70210 [2] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-174-100:70019:70212 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-174-100:70019:70212 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-174-100:70022:70205 [3] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-174-186:375438:375628 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-174-186:375443:375626 [5] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-174-186:375440:375630 [2] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-174-186:375438:375628 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-186:375441:375629 [3] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-174-186:375438:375628 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-174-186:375438:375628 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-174-186:375439:375624 [1] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-186:375438:375628 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-174-186:375445:375627 [7] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-174-186:375438:375628 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-186:375442:375623 [4] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-174-186:375438:375628 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-174-186:375438:375628 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-174-186:375444:375625 [6] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-174-100:70019:70019 [0] NCCL INFO comm 0x8ee4bc0 rank 0 nranks 16 cudaDev 0 busId 53000 - Abort COMPLETE +ip-26-0-174-100:70019:70438 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-186:375438:375438 [0] NCCL INFO comm 0x7668500 rank 8 nranks 16 cudaDev 0 busId 53000 - Abort COMPLETE +ip-26-0-174-186:375438:375857 [0] NCCL INFO [Service thread] Connection closed by localRank 0 + +ip-26-0-174-100:70020:70211 [1] send_close_deferred:5973 NCCL WARN NET/OFI Attempt to call send_close_deferred with outstanding requests! +ip-26-0-174-100:70020:70211 [1] NCCL INFO transport/net.cc:858 -> 3 +ip-26-0-174-100:70020:70211 [1] NCCL INFO proxy.cc:963 -> 3 +ip-26-0-174-100:70020:70211 [1] NCCL INFO proxy.cc:979 -> 3 +ip-26-0-174-186:375445:375445 [7] NCCL INFO comm 0x8b667b0 rank 15 nranks 16 cudaDev 7 busId ca000 - Abort COMPLETE +ip-26-0-174-186:375438:375857 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-174-186:375445:375851 [7] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-174-186:375442:375442 [4] NCCL INFO comm 0x7c1da70 rank 12 nranks 16 cudaDev 4 busId 97000 - Abort COMPLETE +ip-26-0-174-100:70023:70023 [4] NCCL INFO comm 0x7c1af20 rank 4 nranks 16 cudaDev 4 busId 97000 - Abort COMPLETE +ip-26-0-174-186:375438:375857 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-174-186:375442:375852 [4] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-174-100:70019:70438 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-174-100:70023:70436 [4] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-174-100:70026:70026 [7] NCCL INFO comm 0x8cb6e20 rank 7 nranks 16 cudaDev 7 busId ca000 - Abort COMPLETE +ip-26-0-174-100:70019:70438 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-174-100:70026:70441 [7] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-174-186:375444:375444 [6] NCCL INFO comm 0x8397040 rank 14 nranks 16 cudaDev 6 busId b9000 - Abort COMPLETE +ip-26-0-174-186:375438:375857 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-174-186:375444:375856 [6] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-174-186:375443:375443 [5] NCCL INFO comm 0x7220020 rank 13 nranks 16 cudaDev 5 busId a8000 - Abort COMPLETE +ip-26-0-174-186:375438:375857 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-174-186:375443:375850 [5] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-174-100:70020:70020 [1] NCCL INFO comm 0x89ba7a0 rank 1 nranks 16 cudaDev 1 busId 64000 - Abort COMPLETE +ip-26-0-174-100:70019:70438 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70020:70434 [1] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70025:70025 [6] NCCL INFO comm 0x9010ee0 rank 6 nranks 16 cudaDev 6 busId b9000 - Abort COMPLETE +ip-26-0-174-100:70019:70438 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-174-100:70025:70440 [6] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-174-100:70024:70024 [5] NCCL INFO comm 0x84418b0 rank 5 nranks 16 cudaDev 5 busId a8000 - Abort COMPLETE +ip-26-0-174-100:70019:70438 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-174-100:70024:70439 [5] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-174-186:375439:375439 [1] NCCL INFO comm 0x74129a0 rank 9 nranks 16 cudaDev 1 busId 64000 - Abort COMPLETE +ip-26-0-174-186:375438:375857 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-186:375439:375854 [1] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70022:70022 [3] NCCL INFO comm 0x7931080 rank 3 nranks 16 cudaDev 3 busId 86000 - Abort COMPLETE +ip-26-0-174-100:70019:70438 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-174-100:70022:70435 [3] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-174-186:375441:375441 [3] NCCL INFO comm 0x79cca40 rank 11 nranks 16 cudaDev 3 busId 86000 - Abort COMPLETE +ip-26-0-174-186:375438:375857 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-174-186:375441:375853 [3] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-174-100:70021:70021 [2] NCCL INFO comm 0x7f291a0 rank 2 nranks 16 cudaDev 2 busId 75000 - Abort COMPLETE +ip-26-0-174-100:70019:70438 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-174-100:70021:70437 [2] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-174-186:375440:375440 [2] NCCL INFO comm 0x7e24920 rank 10 nranks 16 cudaDev 2 busId 75000 - Abort COMPLETE +ip-26-0-174-186:375438:375857 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-174-186:375440:375855 [2] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-174-100:70026:70026 [7] NCCL INFO comm 0xe334c20 rank 7 nranks 16 cudaDev 7 busId ca000 - Abort COMPLETE +ip-26-0-174-100:70026:70331 [7] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-100:70023:70023 [4] NCCL INFO comm 0xd299a00 rank 4 nranks 16 cudaDev 4 busId 97000 - Abort COMPLETE +ip-26-0-174-100:70025:70025 [6] NCCL INFO comm 0xe68f630 rank 6 nranks 16 cudaDev 6 busId b9000 - Abort COMPLETE +ip-26-0-174-186:375444:375444 [6] NCCL INFO comm 0xda161f0 rank 14 nranks 16 cudaDev 6 busId b9000 - Abort COMPLETE +ip-26-0-174-186:375442:375442 [4] NCCL INFO comm 0xd29e280 rank 12 nranks 16 cudaDev 4 busId 97000 - Abort COMPLETE +ip-26-0-174-100:70023:70293 [4] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-186:375445:375445 [7] NCCL INFO comm 0x100cc140 rank 15 nranks 16 cudaDev 7 busId ca000 - Abort COMPLETE +ip-26-0-174-100:70025:70314 [6] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-186:375442:375724 [4] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-186:375444:375743 [6] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-186:375438:375438 [0] NCCL INFO comm 0xcce58f0 rank 8 nranks 16 cudaDev 0 busId 53000 - Abort COMPLETE +ip-26-0-174-100:70024:70024 [5] NCCL INFO comm 0xdac0e20 rank 5 nranks 16 cudaDev 5 busId a8000 - Abort COMPLETE +ip-26-0-174-186:375445:375742 [7] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70019:70019 [0] NCCL INFO comm 0xe70c7c0 rank 0 nranks 16 cudaDev 0 busId 53000 - Abort COMPLETE +ip-26-0-174-100:70024:70304 [5] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-186:375438:375682 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-100:70019:70253 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-100:70020:70020 [1] NCCL INFO comm 0xe0398d0 rank 1 nranks 16 cudaDev 1 busId 64000 - Abort COMPLETE +ip-26-0-174-100:70020:70262 [1] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-100:70022:70022 [3] NCCL INFO comm 0xcfb0380 rank 3 nranks 16 cudaDev 3 busId 86000 - Abort COMPLETE +ip-26-0-174-186:375441:375441 [3] NCCL INFO comm 0xd04bab0 rank 11 nranks 16 cudaDev 3 busId 86000 - Abort COMPLETE +ip-26-0-174-100:70022:70285 [3] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-186:375441:375714 [3] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-186:375443:375443 [5] NCCL INFO comm 0xc89f1d0 rank 13 nranks 16 cudaDev 5 busId a8000 - Abort COMPLETE +ip-26-0-174-186:375442:375442 [4] NCCL INFO comm 0xc2d2b90 rank 0 nranks 2 cudaDev 4 busId 97000 - Abort COMPLETE +ip-26-0-174-186:375442:375713 [4] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-186:375440:375440 [2] NCCL INFO comm 0xd4a3930 rank 10 nranks 16 cudaDev 2 busId 75000 - Abort COMPLETE +ip-26-0-174-186:375444:375444 [6] NCCL INFO comm 0xca4b1b0 rank 0 nranks 2 cudaDev 6 busId b9000 - Abort COMPLETE +ip-26-0-174-186:375444:375734 [6] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-186:375443:375733 [5] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-186:375445:375445 [7] NCCL INFO comm 0x8f1c990 rank 1 nranks 2 cudaDev 7 busId ca000 - Abort COMPLETE +ip-26-0-174-186:375440:375703 [2] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-186:375438:375438 [0] NCCL INFO comm 0xbd1b320 rank 0 nranks 2 cudaDev 0 busId 53000 - Abort COMPLETE +ip-26-0-174-186:375438:375676 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-100:70023:70023 [4] NCCL INFO comm 0xc2cebe0 rank 0 nranks 2 cudaDev 4 busId 97000 - Abort COMPLETE +ip-26-0-174-100:70023:70284 [4] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70025:70025 [6] NCCL INFO comm 0xd6c4820 rank 0 nranks 2 cudaDev 6 busId b9000 - Abort COMPLETE +ip-26-0-174-100:70025:70303 [6] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70024:70024 [5] NCCL INFO comm 0xcaf5e60 rank 0 nranks 2 cudaDev 5 busId a8000 - Abort COMPLETE +ip-26-0-174-100:70024:70294 [5] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70021:70021 [2] NCCL INFO comm 0xd5a7e20 rank 2 nranks 16 cudaDev 2 busId 75000 - Abort COMPLETE +ip-26-0-174-100:70019:70019 [0] NCCL INFO comm 0xbb0bce0 rank 0 nranks 2 cudaDev 0 busId 53000 - Abort COMPLETE +ip-26-0-174-100:70020:70020 [1] NCCL INFO comm 0xd06e860 rank 0 nranks 2 cudaDev 1 busId 64000 - Abort COMPLETE +ip-26-0-174-100:70020:70254 [1] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70021:70274 [2] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-174-100:70022:70022 [3] NCCL INFO comm 0xbfe5b20 rank 0 nranks 2 cudaDev 3 busId 86000 - Abort COMPLETE +ip-26-0-174-100:70022:70275 [3] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-186:375441:375441 [3] NCCL INFO comm 0xc080ed0 rank 0 nranks 2 cudaDev 3 busId 86000 - Abort COMPLETE +ip-26-0-174-186:375441:375704 [3] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-186:375442:375442 [4] NCCL INFO comm 0x7fd40f0 rank 1 nranks 2 cudaDev 4 busId 97000 - Abort COMPLETE +ip-26-0-174-186:375444:375444 [6] NCCL INFO comm 0x874d430 rank 1 nranks 2 cudaDev 6 busId b9000 - Abort COMPLETE +ip-26-0-174-186:375443:375443 [5] NCCL INFO comm 0xb8d45d0 rank 0 nranks 2 cudaDev 5 busId a8000 - Abort COMPLETE +ip-26-0-174-186:375443:375723 [5] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-186:375440:375440 [2] NCCL INFO comm 0xc4d8630 rank 0 nranks 2 cudaDev 2 busId 75000 - Abort COMPLETE +ip-26-0-174-186:375440:375692 [2] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70023:70023 [4] NCCL INFO comm 0x7fd0e50 rank 1 nranks 2 cudaDev 4 busId 97000 - Abort COMPLETE +ip-26-0-174-100:70026:70026 [7] NCCL INFO comm 0xd36aaa0 rank 0 nranks 2 cudaDev 7 busId ca000 - Abort COMPLETE +ip-26-0-174-100:70026:70315 [7] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70025:70025 [6] NCCL INFO comm 0x93c68a0 rank 1 nranks 2 cudaDev 6 busId b9000 - Abort COMPLETE +ip-26-0-174-100:70024:70024 [5] NCCL INFO comm 0x87f7e70 rank 1 nranks 2 cudaDev 5 busId a8000 - Abort COMPLETE +ip-26-0-174-100:70020:70020 [1] NCCL INFO comm 0x8d70cb0 rank 1 nranks 2 cudaDev 1 busId 64000 - Abort COMPLETE +ip-26-0-174-100:70021:70021 [2] NCCL INFO comm 0xc5dcea0 rank 0 nranks 2 cudaDev 2 busId 75000 - Abort COMPLETE +ip-26-0-174-100:70021:70263 [2] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-174-100:70022:70022 [3] NCCL INFO comm 0x7ce7580 rank 1 nranks 2 cudaDev 3 busId 86000 - Abort COMPLETE +ip-26-0-174-186:375441:375441 [3] NCCL INFO comm 0x7d82dc0 rank 1 nranks 2 cudaDev 3 busId 86000 - Abort COMPLETE +ip-26-0-174-186:375443:375443 [5] NCCL INFO comm 0x75d65c0 rank 1 nranks 2 cudaDev 5 busId a8000 - Abort COMPLETE +ip-26-0-174-186:375440:375440 [2] NCCL INFO comm 0x81dab90 rank 1 nranks 2 cudaDev 2 busId 75000 - Abort COMPLETE +ip-26-0-174-100:70026:70026 [7] NCCL INFO comm 0x906d020 rank 1 nranks 2 cudaDev 7 busId ca000 - Abort COMPLETE +ip-26-0-174-100:70021:70021 [2] NCCL INFO comm 0x82df200 rank 1 nranks 2 cudaDev 2 busId 75000 - Abort COMPLETE +ip-26-0-174-186:375438:375438 [0] NCCL INFO comm 0x7a1d780 rank 1 nranks 2 cudaDev 0 busId 53000 - Abort COMPLETE +[2024-12-20 12:07:50,417] torch.distributed.elastic.agent.server.api: [ERROR] Error waiting on exit barrier. Elapsed: 300.116907119751 seconds +[2024-12-20 12:07:50,468] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-174-186.ec2.internal_375361_0' has failed to send a keep-alive heartbeat to the rendezvous '13523084' due to an error of type RendezvousConnectionError. +[2024-12-20 12:07:50,571] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 375439 closing signal SIGTERM +[2024-12-20 12:07:54,398] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-174-186.ec2.internal_375361_0' has failed to shutdown the rendezvous '13523084' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/rendezvous/c10d_rendezvous_backend.py", line 113, in _call_store + return getattr(self._store, store_op)(*args, **kwargs) +RuntimeError: Broken pipe + +The above exception was the direct cause of the following exception: + +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 909, in _invoke_run + num_nodes_waiting = rdzv_handler.num_nodes_waiting() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/rendezvous/dynamic_rendezvous.py", line 1083, in num_nodes_waiting + self._state_holder.sync() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/rendezvous/dynamic_rendezvous.py", line 409, in sync + get_response = self._backend.get_state() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/rendezvous/c10d_rendezvous_backend.py", line 73, in get_state + base64_state: bytes = self._call_store("get", self._key) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/rendezvous/c10d_rendezvous_backend.py", line 115, in _call_store + raise RendezvousConnectionError( +torch.distributed.elastic.rendezvous.api.RendezvousConnectionError: The connection to the C10d store has failed. See inner exception for details. +srun: error: ip-26-0-174-186: task 1: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13523084.0 diff --git a/logs/13767063-bench_1.34G_dp1_tp2_pp8_acc16_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13767063-bench_1.34G_dp1_tp2_pp8_acc16_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..b3d7a9d3355dc89ade97e923ef225f6fbf845d61 --- /dev/null +++ b/logs/13767063-bench_1.34G_dp1_tp2_pp8_acc16_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,5757 @@ ++ '[' -z 13767063 ']' ++ unset FI_PROVIDER ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-170-[143,160]' ++ export 'NODELIST=ip-26-0-170-143 +ip-26-0-170-160' ++ NODELIST='ip-26-0-170-143 +ip-26-0-170-160' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-170-[143,160]' ++ export MASTER_NODE=ip-26-0-170-143 ++ MASTER_NODE=ip-26-0-170-143 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ export NCCL_P2P_LEVEL=LOC ++ NCCL_P2P_LEVEL=LOC ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-170-143' +Master node: ip-26-0-170-143 ++ echo 'All nodes: ip-26-0-170-143 +ip-26-0-170-160' +All nodes: ip-26-0-170-143 +ip-26-0-170-160 ++ echo 'World size: 16' +World size: 16 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13767063 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-170-143:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.34G_dp1_tp2_pp8_acc16_mbs2_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-27 15:53:06,049] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 15:53:06,049] torch.distributed.run: [WARNING] +[2024-12-27 15:53:06,049] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 15:53:06,049] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 15:53:06,049] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 15:53:09,538] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 15:53:09,538] torch.distributed.run: [WARNING] +[2024-12-27 15:53:09,538] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 15:53:09,538] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 15:53:09,538] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-170-143:2067653:2067653 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2067653:2067653 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2067653:2067653 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2067653:2067653 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-143:2067654:2067654 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2067657:2067657 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2067654:2067654 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2067657:2067657 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2067654:2067654 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2067654:2067654 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2067657:2067657 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2067657:2067657 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2102342:2102342 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2102338:2102338 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2102345:2102345 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2102339:2102339 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2102344:2102344 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2102341:2102341 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2102343:2102343 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2102340:2102340 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2067658:2067658 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2067655:2067655 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2067659:2067659 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2067656:2067656 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2067658:2067658 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2067658:2067658 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2067658:2067658 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2067660:2067660 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2067655:2067655 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2067655:2067655 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2067655:2067655 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2067659:2067659 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2067656:2067656 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2067659:2067659 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2067659:2067659 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2067660:2067660 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2067656:2067656 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2067656:2067656 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2067660:2067660 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2067660:2067660 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2102341:2102341 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2102342:2102342 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2102345:2102345 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2102344:2102344 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2102340:2102340 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2102343:2102343 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2102338:2102338 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2102339:2102339 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2102343:2102343 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2102344:2102344 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2102339:2102339 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2102341:2102341 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2102345:2102345 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2102343:2102343 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2102344:2102344 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2102339:2102339 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2102345:2102345 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2102341:2102341 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2102342:2102342 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2102342:2102342 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2102338:2102338 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2102338:2102338 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2102340:2102340 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2102340:2102340 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/205 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO comm 0x945ccd0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO comm 0x9919d80 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO comm 0x8915d70 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO comm 0xa16ef10 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO comm 0x96f8980 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO comm 0x951df00 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO comm 0x907c020 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO comm 0xa256590 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO comm 0x9ae7290 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO comm 0xa3c1710 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO comm 0x8e7e9b0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO comm 0x9e46280 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO comm 0x9e63c20 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO comm 0x96486c0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO comm 0x87b79d0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO comm 0xa219ae0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd207b0bafb6d8d84 - Init START +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 7/-1/-1->15->-1 [3] 7/-1/-1->15->-1 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Trees [0] 8/-1/-1->0->-1 [1] 8/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Trees [0] 4/12/-1->8->0 [1] 4/12/-1->8->0 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102345:2102508 [7] NCCL INFO comm 0x9ae7290 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067653:2067822 [0] NCCL INFO comm 0xa256590 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102340:2102510 [2] NCCL INFO comm 0x87b79d0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-160:2102344:2102512 [6] NCCL INFO comm 0xa3c1710 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067659:2067825 [6] NCCL INFO comm 0x951df00 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-143:2067655:2067826 [2] NCCL INFO comm 0xa16ef10 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067654:2067824 [1] NCCL INFO comm 0x907c020 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-160:2102343:2102509 [5] NCCL INFO comm 0xa219ae0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-160:2102339:2102505 [1] NCCL INFO comm 0x96486c0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-143:2067658:2067827 [5] NCCL INFO comm 0x9919d80 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102341:2102506 [3] NCCL INFO comm 0x9e63c20 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-143:2067656:2067828 [3] NCCL INFO comm 0x96f8980 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067657:2067823 [4] NCCL INFO comm 0x8915d70 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-143:2067660:2067829 [7] NCCL INFO comm 0x945ccd0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-160:2102342:2102511 [4] NCCL INFO comm 0x9e46280 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-160:2102338:2102507 [0] NCCL INFO comm 0x8e7e9b0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd207b0bafb6d8d84 - Init COMPLETE +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO comm 0x9d20100 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO comm 0x96957a0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO comm 0x9b528e0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO comm 0x97570e0 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO comm 0xa492490 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO comm 0x92b4be0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO comm 0xa3a8220 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO comm 0x8b4ece0 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO comm 0x9931fe0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO comm 0xa5fa500 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO comm 0xa452cc0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO comm 0xa07ee40 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO comm 0xa09c4f0 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO comm 0x89f0ae0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO comm 0x90b6ac0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO comm 0x9880f80 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4eea597bedfc78e5 - Init START +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 7/-1/-1->15->-1 [3] 7/-1/-1->15->-1 +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Trees [0] 8/-1/-1->0->-1 [1] 8/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Trees [0] 4/12/-1->8->0 [1] 4/12/-1->8->0 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102345:2102709 [7] NCCL INFO comm 0x9d20100 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067653:2068022 [0] NCCL INFO comm 0xa492490 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102340:2102706 [2] NCCL INFO comm 0x89f0ae0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-160:2102344:2102705 [6] NCCL INFO comm 0xa5fa500 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-143:2067659:2068023 [6] NCCL INFO comm 0x97570e0 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067655:2068026 [2] NCCL INFO comm 0xa3a8220 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067654:2068024 [1] NCCL INFO comm 0x92b4be0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-143:2067658:2068029 [5] NCCL INFO comm 0x9b528e0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102343:2102704 [5] NCCL INFO comm 0xa452cc0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067656:2068028 [3] NCCL INFO comm 0x9931fe0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102341:2102707 [3] NCCL INFO comm 0xa09c4f0 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-160:2102342:2102703 [4] NCCL INFO comm 0xa07ee40 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-143:2067657:2068025 [4] NCCL INFO comm 0x8b4ece0 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-143:2067660:2068027 [7] NCCL INFO comm 0x96957a0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102339:2102710 [1] NCCL INFO comm 0x9880f80 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4eea597bedfc78e5 - Init COMPLETE +ip-26-0-170-160:2102338:2102708 [0] NCCL INFO comm 0x90b6ac0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4eea597bedfc78e5 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO comm 0x8a04f40 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7421b120f7dde3f7 - Init START +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO comm 0xa467580 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7421b120f7dde3f7 - Init START +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO comm 0xa0b0720 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7421b120f7dde3f7 - Init START +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO comm 0x9895b90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7421b120f7dde3f7 - Init START +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO comm 0x90cc180 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7421b120f7dde3f7 - Init START +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO comm 0x9d33b50 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7421b120f7dde3f7 - Init START +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO comm 0xa60e850 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7421b120f7dde3f7 - Init START +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO comm 0xa093650 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7421b120f7dde3f7 - Init START +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO comm 0xa4a6ac0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc495b6685f6acbcc - Init START +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO comm 0x92c8650 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc495b6685f6acbcc - Init START +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO comm 0x9945b20 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc495b6685f6acbcc - Init START +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO comm 0xa3bbe40 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc495b6685f6acbcc - Init START +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO comm 0x8b626a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc495b6685f6acbcc - Init START +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO comm 0x976b300 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc495b6685f6acbcc - Init START +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO comm 0x96a9980 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc495b6685f6acbcc - Init START +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO comm 0x9b66440 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc495b6685f6acbcc - Init START +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102345:2102744 [7] NCCL INFO comm 0x9d33b50 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7421b120f7dde3f7 - Init COMPLETE +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102338:2102737 [0] NCCL INFO comm 0x90cc180 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7421b120f7dde3f7 - Init COMPLETE +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102344:2102739 [6] NCCL INFO comm 0xa60e850 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7421b120f7dde3f7 - Init COMPLETE +ip-26-0-170-160:2102339:2102743 [1] NCCL INFO comm 0x9895b90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7421b120f7dde3f7 - Init COMPLETE +ip-26-0-170-160:2102340:2102740 [2] NCCL INFO comm 0x8a04f40 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7421b120f7dde3f7 - Init COMPLETE +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102341:2102742 [3] NCCL INFO comm 0xa0b0720 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7421b120f7dde3f7 - Init COMPLETE +ip-26-0-170-160:2102343:2102741 [5] NCCL INFO comm 0xa467580 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7421b120f7dde3f7 - Init COMPLETE +ip-26-0-170-160:2102342:2102738 [4] NCCL INFO comm 0xa093650 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7421b120f7dde3f7 - Init COMPLETE +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067660:2068057 [7] NCCL INFO comm 0x96a9980 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc495b6685f6acbcc - Init COMPLETE +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067653:2068055 [0] NCCL INFO comm 0xa4a6ac0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc495b6685f6acbcc - Init COMPLETE +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067659:2068056 [6] NCCL INFO comm 0x976b300 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc495b6685f6acbcc - Init COMPLETE +ip-26-0-170-143:2067655:2068058 [2] NCCL INFO comm 0xa3bbe40 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc495b6685f6acbcc - Init COMPLETE +ip-26-0-170-143:2067654:2068060 [1] NCCL INFO comm 0x92c8650 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc495b6685f6acbcc - Init COMPLETE +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067656:2068061 [3] NCCL INFO comm 0x9945b20 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc495b6685f6acbcc - Init COMPLETE +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067658:2068059 [5] NCCL INFO comm 0x9b66440 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc495b6685f6acbcc - Init COMPLETE +ip-26-0-170-143:2067657:2068062 [4] NCCL INFO comm 0x8b626a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc495b6685f6acbcc - Init COMPLETE +12/27/2024 15:53:49 [INFO|DP=0|PP=6|TP=0|ip-26-0-170-160]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=4|TP=1|ip-26-0-170-160]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=5|TP=0|ip-26-0-170-160]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=5|TP=1|ip-26-0-170-160]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=4|TP=0|ip-26-0-170-160]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=6|TP=1|ip-26-0-170-160]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=7|TP=1|ip-26-0-170-160]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Config: +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Config(general=GeneralArgs(project='debug', +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: run='1.34G_dp1_tp2_pp8_acc16_mbs2_seq4096_zero1_tpmodeRED_vocab131k', +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: seed=42, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: step=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: consumed_train_samples=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2_pp.csv'), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: ignore_sanity_checks=True), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: parallelism=ParallelismArgs(dp=1, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pp=8, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp=2, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pp_engine=, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp_mode=, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp_linear_async_communication=True, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: recompute_layer=False, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp_recompute_allgather=True, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: expert_parallel_size=1), +12/27/2024 15:53:49 [INFO|DP=0|PP=3|TP=1|ip-26-0-170-143]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: eos_token_id=0, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_act='silu', +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_size=2048, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: initializer_range=0.02, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: intermediate_size=8192, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: is_llama_config=True, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: max_position_embeddings=4096, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_attention_heads=32, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_hidden_layers=16, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_key_value_heads=32, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pad_token_id=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pretraining_tp=1, +12/27/2024 15:53:49 [INFO|DP=0|PP=2|TP=1|ip-26-0-170-143]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rms_norm_eps=1e-05, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_scaling=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_theta=10000.0, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_interleaved=False, +12/27/2024 15:53:49 [INFO|DP=0|PP=3|TP=0|ip-26-0-170-143]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tie_word_embeddings=True, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: use_cache=True, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: vocab_size=131072), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: init_method=RandomInit(std=0.02), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: dtype=torch.bfloat16, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: make_vocab_size_divisible_by=1, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: ddp_bucket_cap_mb=25), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokenizer_revision=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokenizer_max_length=None), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: checkpoint_interval=10000, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: save_initial_state=False, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: save_final_state=False, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: resume_checkpoint_path=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: checkpoints_path_is_shared_file_system=False), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: logging=LoggingArgs(log_level='info', +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: log_level_replica='info', +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: iteration_step_info_interval=1), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokens=TokensArgs(sequence_length=4096, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: train_steps=100, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: micro_batch_size=2, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: batch_accumulation_per_replica=16, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: val_check_interval=100, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: limit_val_batches=0, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: limit_test_batches=0), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: adam_beta1=0.9, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: adam_beta2=0.95, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: torch_adam_is_fused=True, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: name='adamW'), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: zero_stage=1, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: weight_decay=0.01, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: clip_grad=1.0, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: accumulate_grad_in_fp32=True, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_warmup_steps=2, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_warmup_style='linear', +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_decay_style='cosine', +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_decay_steps=13, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_decay_starting_step=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: min_decay_lr=1e-05)), +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: start_training_step=1, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: data=DataArgs(dataset=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: seed=42, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_loading_workers=1))], +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: profiler=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lighteval=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: s3_upload=None) +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=1|ip-26-0-170-143]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Model Config: +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: LlamaConfig(bos_token_id=0, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: eos_token_id=0, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_act='silu', +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_size=2048, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: initializer_range=0.02, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: intermediate_size=8192, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: is_llama_config=True, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: max_position_embeddings=4096, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_attention_heads=32, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_hidden_layers=16, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_key_value_heads=32, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pad_token_id=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pretraining_tp=1, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rms_norm_eps=1e-05, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_scaling=None, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_theta=10000.0, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_interleaved=False, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tie_word_embeddings=True, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: use_cache=True, +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: vocab_size=131072) +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Building model.. +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Initialize RoPE Theta = 10000.0 +12/27/2024 15:53:49 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-143]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=2|TP=0|ip-26-0-170-143]: [INIT] P2P Communication: enabled, Device: cuda, Group: +12/27/2024 15:53:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Setting PP block ranks... +12/27/2024 15:53:49 [INFO|DP=0|PP=1|TP=1|ip-26-0-170-143]: [INIT] P2P Communication: enabled, Device: cuda, Group: +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO comm 0xc179c60 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x770a6cba3a705e95 - Init START +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO comm 0xbda5ae0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x770a6cba3a705e95 - Init START +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO comm 0xbdc16a0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x88dd328cab6f992c - Init START +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO comm 0xa716120 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x88dd328cab6f992c - Init START +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO comm 0xadde760 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x24298a447b6931f0 - Init START +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO comm 0xb5a75a0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x24298a447b6931f0 - Init START +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO comm 0xb659220 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x282ac160b117c38d - Init START +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO comm 0xc0d0630 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x282ac160b117c38d - Init START +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO comm 0xb47c960 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x72eea4401434656b - Init START +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO comm 0xb3bc220 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x72eea4401434656b - Init START +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO comm 0xa874ce0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc754e2c88a7b0dc7 - Init START +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO comm 0xb876fe0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc754e2c88a7b0dc7 - Init START +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO comm 0xc37e1c0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd4e9806796d5452a - Init START +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO comm 0xb19bde0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd4e9806796d5452a - Init START +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO comm 0xc31e340 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7b0107281f2d0a25 - Init START +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO comm 0xba43f20 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7b0107281f2d0a25 - Init START +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067660:2068100 [7] NCCL INFO comm 0xb3bc220 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x72eea4401434656b - Init COMPLETE +ip-26-0-170-143:2067659:2068098 [6] NCCL INFO comm 0xb47c960 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x72eea4401434656b - Init COMPLETE +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067656:2068103 [3] NCCL INFO comm 0xb659220 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x282ac160b117c38d - Init COMPLETE +ip-26-0-170-143:2067655:2068101 [2] NCCL INFO comm 0xc0d0630 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x282ac160b117c38d - Init COMPLETE +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067654:2068112 [1] NCCL INFO comm 0xb19bde0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd4e9806796d5452a - Init COMPLETE +ip-26-0-170-143:2067653:2068111 [0] NCCL INFO comm 0xc37e1c0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd4e9806796d5452a - Init COMPLETE +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067657:2068099 [4] NCCL INFO comm 0xa874ce0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc754e2c88a7b0dc7 - Init COMPLETE +ip-26-0-170-143:2067658:2068102 [5] NCCL INFO comm 0xb876fe0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc754e2c88a7b0dc7 - Init COMPLETE +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102343:2102779 [5] NCCL INFO comm 0xc179c60 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x770a6cba3a705e95 - Init COMPLETE +ip-26-0-170-160:2102342:2102777 [4] NCCL INFO comm 0xbda5ae0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x770a6cba3a705e95 - Init COMPLETE +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102341:2102783 [3] NCCL INFO comm 0xbdc16a0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x88dd328cab6f992c - Init COMPLETE +ip-26-0-170-160:2102340:2102780 [2] NCCL INFO comm 0xa716120 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x88dd328cab6f992c - Init COMPLETE +ip-26-0-170-160:2102339:2102786 [1] NCCL INFO comm 0xb5a75a0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x24298a447b6931f0 - Init COMPLETE +ip-26-0-170-160:2102344:2102792 [6] NCCL INFO comm 0xc31e340 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7b0107281f2d0a25 - Init COMPLETE +ip-26-0-170-160:2102338:2102782 [0] NCCL INFO comm 0xadde760 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x24298a447b6931f0 - Init COMPLETE +ip-26-0-170-160:2102345:2102793 [7] NCCL INFO comm 0xba43f20 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7b0107281f2d0a25 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO comm 0xa8879a0 rank 2 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2ba7d2075f58e9d2 - Init START +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO comm 0xb48f620 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2ba7d2075f58e9d2 - Init START +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO comm 0xc331000 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2ba7d2075f58e9d2 - Init START +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO comm 0xc390e80 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2ba7d2075f58e9d2 - Init START +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO comm 0xc0e32f0 rank 1 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2ba7d2075f58e9d2 - Init START +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO comm 0xadf1420 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2ba7d2075f58e9d2 - Init START +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO comm 0xbdb87a0 rank 6 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2ba7d2075f58e9d2 - Init START +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO comm 0xa728de0 rank 5 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2ba7d2075f58e9d2 - Init START +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO comm 0xb3ceee0 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x15f073aa4f367550 - Init START +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO comm 0xb889ca0 rank 2 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x15f073aa4f367550 - Init START +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO comm 0xb66bee0 rank 1 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x15f073aa4f367550 - Init START +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO comm 0xc18c920 rank 6 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x15f073aa4f367550 - Init START +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO comm 0xb1aeaa0 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x15f073aa4f367550 - Init START +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO comm 0xba56be0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x15f073aa4f367550 - Init START +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO comm 0xbdd4360 rank 5 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x15f073aa4f367550 - Init START +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO comm 0xb5ba260 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x15f073aa4f367550 - Init START +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 00/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 01/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 02/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 03/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 00/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 01/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 02/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 03/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 00/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 00/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 01/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 01/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 02/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 02/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 00/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 03/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 00/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 03/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 01/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 00/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 02/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 01/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 01/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 02/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 03/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 00/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 02/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 00/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 03/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 03/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 01/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 00/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 01/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 02/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 02/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 01/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 03/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 00/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 00/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 03/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 02/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 00/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 01/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 03/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 01/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 01/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 02/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 02/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 02/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 03/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 03/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 03/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 00/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 01/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 02/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 03/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 00/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 00/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 01/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 02/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 01/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 03/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 02/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 00/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 03/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 01/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 00/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 00/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 02/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 01/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 01/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 03/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 02/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 03/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 02/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 03/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 00/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 01/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 02/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 03/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 00/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 01/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 02/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 03/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 00/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 01/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 02/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 03/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 00/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 01/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 02/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 03/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 00/0 : 4[0] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 01/0 : 4[0] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 02/0 : 3[7] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 03/0 : 3[7] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 02/0 : 1[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 03/0 : 1[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 02/0 : 3[6] -> 5[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 03/0 : 3[6] -> 5[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 00/0 : 4[1] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 01/0 : 4[1] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 00/0 : 2[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 01/0 : 2[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 02/0 : 3[6] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 03/0 : 3[6] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 02/0 : 5[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 03/0 : 5[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 00/0 : 2[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 01/0 : 2[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 02/0 : 3[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 03/0 : 3[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 02/0 : 1[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 03/0 : 1[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 02/0 : 3[7] -> 5[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 03/0 : 3[7] -> 5[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 00/0 : 7[6] -> 6[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 02/0 : 5[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 03/0 : 5[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Channel 01/0 : 7[6] -> 6[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 02/0 : 5[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 03/0 : 5[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 00/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 00/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 01/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 02/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 03/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 02/0 : 3[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 03/0 : 3[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 01/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 02/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 03/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 02/0 : 5[2] -> 4[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Channel 03/0 : 5[2] -> 4[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 02/0 : 5[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 03/0 : 5[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 00/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 00/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 01/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 01/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 00/0 : 2[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 02/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 02/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 01/0 : 2[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 03/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 03/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 00/0 : 4[1] -> 6[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 02/0 : 5[3] -> 4[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 01/0 : 4[1] -> 6[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Channel 03/0 : 5[3] -> 4[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 00/0 : 4[1] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 00/0 : 6[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 01/0 : 4[1] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 01/0 : 6[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 00/0 : 6[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 01/0 : 6[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 00/0 : 4[1] -> 2[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 01/0 : 4[1] -> 2[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 00/0 : 7[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 01/0 : 7[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 00/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 00/0 : 3[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 02/0 : 5[3] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 01/0 : 3[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 01/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Channel 03/0 : 5[3] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 02/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Channel 03/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 00/0 : 2[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 01/0 : 2[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 00/0 : 4[0] -> 6[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 01/0 : 4[0] -> 6[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 00/0 : 4[0] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 00/0 : 6[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 01/0 : 4[0] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 01/0 : 6[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 00/0 : 6[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 01/0 : 6[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 00/0 : 4[0] -> 2[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 01/0 : 4[0] -> 2[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 00/0 : 7[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 01/0 : 7[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 02/0 : 5[2] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 00/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 00/0 : 3[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 01/0 : 3[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 01/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Channel 03/0 : 5[2] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 02/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Channel 03/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067654:2068137 [1] NCCL INFO comm 0xb1aeaa0 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x15f073aa4f367550 - Init COMPLETE +ip-26-0-170-160:2102344:2102812 [6] NCCL INFO comm 0xc331000 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2ba7d2075f58e9d2 - Init COMPLETE +ip-26-0-170-160:2102345:2102818 [7] NCCL INFO comm 0xba56be0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x15f073aa4f367550 - Init COMPLETE +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067653:2068132 [0] NCCL INFO comm 0xc390e80 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2ba7d2075f58e9d2 - Init COMPLETE +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067656:2068139 [3] NCCL INFO comm 0xb66bee0 rank 1 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x15f073aa4f367550 - Init COMPLETE +ip-26-0-170-160:2102341:2102817 [3] NCCL INFO comm 0xbdd4360 rank 5 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x15f073aa4f367550 - Init COMPLETE +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067658:2068140 [5] NCCL INFO comm 0xb889ca0 rank 2 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x15f073aa4f367550 - Init COMPLETE +ip-26-0-170-160:2102343:2102816 [5] NCCL INFO comm 0xc18c920 rank 6 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x15f073aa4f367550 - Init COMPLETE +ip-26-0-170-143:2067655:2068136 [2] NCCL INFO comm 0xc0e32f0 rank 1 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2ba7d2075f58e9d2 - Init COMPLETE +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102342:2102815 [4] NCCL INFO comm 0xbdb87a0 rank 6 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2ba7d2075f58e9d2 - Init COMPLETE +ip-26-0-170-143:2067657:2068133 [4] NCCL INFO comm 0xa8879a0 rank 2 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2ba7d2075f58e9d2 - Init COMPLETE +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067660:2068138 [7] NCCL INFO comm 0xb3ceee0 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x15f073aa4f367550 - Init COMPLETE +ip-26-0-170-143:2067659:2068134 [6] NCCL INFO comm 0xb48f620 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2ba7d2075f58e9d2 - Init COMPLETE +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102339:2102819 [1] NCCL INFO comm 0xb5ba260 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x15f073aa4f367550 - Init COMPLETE +ip-26-0-170-160:2102340:2102813 [2] NCCL INFO comm 0xa728de0 rank 5 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2ba7d2075f58e9d2 - Init COMPLETE +ip-26-0-170-160:2102338:2102814 [0] NCCL INFO comm 0xadf1420 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2ba7d2075f58e9d2 - Init COMPLETE +12/27/2024 15:53:55 [INFO|DP=0|PP=0|TP=1|ip-26-0-170-143]: Local number of parameters: 168M (320.01MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=0|TP=1|ip-26-0-170-143]: [After model building] Memory usage: 320.02MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO comm 0xb1c3960 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x229e1f2cd2d645de - Init START +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +12/27/2024 15:53:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Total number of parameters: 1.61G (3072.26MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Local number of parameters: 168M (320.01MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [After model building] Memory usage: 320.02MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/27/2024 15:53:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: No checkpoint path provided. +12/27/2024 15:53:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Parametrizing model parameters using StandardParametrizator +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO comm 0xc3a4170 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4969bbaeb8723ca9 - Init START +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +12/27/2024 15:53:55 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: Local number of parameters: 168M (320.01MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=4|TP=0|ip-26-0-170-160]: Local number of parameters: 101M (192.02MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=6|TP=0|ip-26-0-170-160]: Local number of parameters: 67.1M (128.02MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=5|TP=0|ip-26-0-170-160]: Local number of parameters: 67.1M (128.02MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 320.02MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/27/2024 15:53:55 [INFO|DP=0|PP=4|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 192.04MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/27/2024 15:53:55 [INFO|DP=0|PP=6|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 128.03MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/27/2024 15:53:55 [INFO|DP=0|PP=5|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 128.03MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/27/2024 15:53:55 [INFO|DP=0|PP=3|TP=0|ip-26-0-170-143]: Local number of parameters: 67.1M (128.02MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=2|TP=0|ip-26-0-170-143]: Local number of parameters: 67.1M (128.02MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-143]: Local number of parameters: 101M (192.02MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=3|TP=0|ip-26-0-170-143]: [After model building] Memory usage: 128.03MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/27/2024 15:53:55 [INFO|DP=0|PP=2|TP=0|ip-26-0-170-143]: [After model building] Memory usage: 128.03MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/27/2024 15:53:55 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-143]: [After model building] Memory usage: 192.04MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO comm 0xc505460 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x65b4a2ada325f51c - Init START +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO comm 0xbf8a920 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x271e4abdd1cfb9c4 - Init START +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO comm 0xafc31f0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9dca418b2c7ff36a - Init START +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO comm 0xa8fb800 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x197e430a39143a24 - Init START +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO comm 0xb662e00 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcb18ee1617cae821 - Init START +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +12/27/2024 15:53:55 [INFO|DP=0|PP=4|TP=1|ip-26-0-170-160]: Local number of parameters: 101M (192.02MiB) +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO comm 0xaa597a0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x47680d284fa72234 - Init START +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +12/27/2024 15:53:55 [INFO|DP=0|PP=4|TP=1|ip-26-0-170-160]: [After model building] Memory usage: 192.04MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/27/2024 15:53:55 [INFO|DP=0|PP=2|TP=1|ip-26-0-170-143]: Local number of parameters: 67.1M (128.02MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=2|TP=1|ip-26-0-170-143]: [After model building] Memory usage: 128.03MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Using network Libfabric +12/27/2024 15:53:55 [INFO|DP=0|PP=1|TP=1|ip-26-0-170-143]: Local number of parameters: 101M (192.02MiB) +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO comm 0xc2b61c0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6938f954443389e2 - Init START +12/27/2024 15:53:55 [INFO|DP=0|PP=3|TP=1|ip-26-0-170-143]: Local number of parameters: 67.1M (128.02MiB) +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +12/27/2024 15:53:55 [INFO|DP=0|PP=1|TP=1|ip-26-0-170-143]: [After model building] Memory usage: 192.04MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/27/2024 15:53:55 [INFO|DP=0|PP=3|TP=1|ip-26-0-170-143]: [After model building] Memory usage: 128.03MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +12/27/2024 15:53:55 [INFO|DP=0|PP=6|TP=1|ip-26-0-170-160]: Local number of parameters: 67.1M (128.02MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=6|TP=1|ip-26-0-170-160]: [After model building] Memory usage: 128.03MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/27/2024 15:53:55 [INFO|DP=0|PP=5|TP=1|ip-26-0-170-160]: Local number of parameters: 67.1M (128.02MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=7|TP=1|ip-26-0-170-160]: Local number of parameters: 168M (320.01MiB) +12/27/2024 15:53:55 [INFO|DP=0|PP=5|TP=1|ip-26-0-170-160]: [After model building] Memory usage: 128.03MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Using network Libfabric +12/27/2024 15:53:55 [INFO|DP=0|PP=7|TP=1|ip-26-0-170-160]: [After model building] Memory usage: 320.02MiB. Peak allocated: 5504.00MiB Peak reserved: 20866.00MiB +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO comm 0xba5cd60 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x904773b230bf07b0 - Init START +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO comm 0xb78bef0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc6feda6c7eb824a3 - Init START +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO comm 0xb5a2240 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf3edff7cd05abda0 - Init START +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO comm 0xb83e890 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x444f323fbe74a2d7 - Init START +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO comm 0xc35e840 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x85de4387989e493c - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO comm 0xbfa5a00 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x23c4cf9bb1a261be - Init START +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO comm 0xbc2a610 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfb3281d42640e259 - Init START +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067654:2068162 [1] NCCL INFO comm 0xb1c3960 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x229e1f2cd2d645de - Init COMPLETE +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2067658:2068173 [5] NCCL INFO comm 0xba5cd60 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x904773b230bf07b0 - Init COMPLETE +ip-26-0-170-143:2067653:2068165 [0] NCCL INFO comm 0xc3a4170 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4969bbaeb8723ca9 - Init COMPLETE +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:2102342:2102844 [4] NCCL INFO comm 0xbf8a920 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x271e4abdd1cfb9c4 - Init COMPLETE +ip-26-0-170-143:2067655:2068171 [2] NCCL INFO comm 0xc2b61c0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6938f954443389e2 - Init COMPLETE +ip-26-0-170-160:2102340:2102848 [2] NCCL INFO comm 0xa8fb800 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x197e430a39143a24 - Init COMPLETE +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067659:2068167 [6] NCCL INFO comm 0xb662e00 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcb18ee1617cae821 - Init COMPLETE +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102846 [0] NCCL INFO comm 0xafc31f0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9dca418b2c7ff36a - Init COMPLETE +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067660:2068175 [7] NCCL INFO comm 0xb5a2240 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf3edff7cd05abda0 - Init COMPLETE +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2067656:2068177 [3] NCCL INFO comm 0xb83e890 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x444f323fbe74a2d7 - Init COMPLETE +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:2102339:2102850 [1] NCCL INFO comm 0xb78bef0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc6feda6c7eb824a3 - Init COMPLETE +ip-26-0-170-143:2067657:2068169 [4] NCCL INFO comm 0xaa597a0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x47680d284fa72234 - Init COMPLETE +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:2102341:2102854 [3] NCCL INFO comm 0xbfa5a00 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x23c4cf9bb1a261be - Init COMPLETE +ip-26-0-170-160:2102343:2102852 [5] NCCL INFO comm 0xc35e840 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x85de4387989e493c - Init COMPLETE +ip-26-0-170-160:2102345:2102856 [7] NCCL INFO comm 0xbc2a610 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfb3281d42640e259 - Init COMPLETE +ip-26-0-170-160:2102344:2102842 [6] NCCL INFO comm 0xc505460 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x65b4a2ada325f51c - Init COMPLETE +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO comm 0xc51f6f0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x43b32381fc2c5158 - Init START +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO comm 0xc3bb9c0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x43b32381fc2c5158 - Init START +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO comm 0xb1db400 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xaf3ea12b6fdb0329 - Init START +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO comm 0xbc44900 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xaf3ea12b6fdb0329 - Init START +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Channel 00/0 : 0[0] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Channel 01/0 : 0[0] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Channel 02/0 : 0[0] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Channel 03/0 : 0[0] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 00/0 : 1[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 01/0 : 1[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 02/0 : 1[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 03/0 : 1[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Channel 00/0 : 0[1] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Channel 01/0 : 0[1] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Channel 02/0 : 0[1] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Channel 03/0 : 0[1] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 00/0 : 1[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 01/0 : 1[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 02/0 : 1[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 03/0 : 1[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067653:2068195 [0] NCCL INFO comm 0xc3bb9c0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x43b32381fc2c5158 - Init COMPLETE +12/27/2024 15:53:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Optimizer Building] Using LearningRateForSP as learning rate +12/27/2024 15:53:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] Size of optimizer params per rank: +12/27/2024 15:53:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] DP Rank 0 has 168M out of 168M (100.00%) params' optimizer states +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067654:2068197 [1] NCCL INFO comm 0xb1db400 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xaf3ea12b6fdb0329 - Init COMPLETE +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102344:2102873 [6] NCCL INFO comm 0xc51f6f0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x43b32381fc2c5158 - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:2102345:2102874 [7] NCCL INFO comm 0xbc44900 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xaf3ea12b6fdb0329 - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/27/2024 15:53:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/27/2024 15:53:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Using dummy data generator +12/27/2024 15:53:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Training Plan] There are 1 training stages +12/27/2024 15:53:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Stage Stable Training Stage] start from step 1 +12/27/2024 15:53:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: +12/27/2024 15:53:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Start training] datetime: 2024-12-27 15:53:56.854634 | mbs: 2 | grad_accum: 16 | global_batch_size: 32 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/27/2024 15:53:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/27/2024 15:53:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 1600.05MiB. Peak allocated 5504.00MiB. Peak reserved: 20866.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO comm 0x1c8e38e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x45797b4a2add66c4 - Init START +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO comm 0xc507c10 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x45797b4a2add66c4 - Init START +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO comm 0xba93140 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb8bb6921d818bc20 - Init START +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO comm 0x1b625b80 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb8bb6921d818bc20 - Init START +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067655:2068215 [2] NCCL INFO comm 0xc507c10 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x45797b4a2add66c4 - Init COMPLETE +ip-26-0-170-143:2067655:2068226 [2] NCCL INFO Channel 02/1 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-143:2067655:2068226 [2] NCCL INFO Channel 03/1 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067653:2068213 [0] NCCL INFO comm 0x1c8e38e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x45797b4a2add66c4 - Init COMPLETE +ip-26-0-170-143:2067653:2068227 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-143:2067653:2068227 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-143:2067656:2068216 [3] NCCL INFO comm 0xba93140 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb8bb6921d818bc20 - Init COMPLETE +ip-26-0-170-143:2067656:2068228 [3] NCCL INFO Channel 02/1 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2067656:2068228 [3] NCCL INFO Channel 03/1 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2067654:2068214 [1] NCCL INFO comm 0x1b625b80 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb8bb6921d818bc20 - Init COMPLETE +ip-26-0-170-143:2067654:2068229 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-143:2067654:2068229 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO comm 0xac13810 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb88431b88f65b221 - Init START +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO comm 0x14e443b0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb88431b88f65b221 - Init START +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO comm 0xbc1be20 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x464fd298d6760e0c - Init START +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO comm 0x148e5840 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x464fd298d6760e0c - Init START +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Channel 00/0 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Channel 01/0 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Channel 02/0 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Channel 03/0 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 00/0 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 01/0 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 02/0 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 03/0 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Channel 00/0 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Channel 01/0 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Channel 02/0 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Channel 03/0 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 00/0 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 01/0 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 02/0 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 03/0 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067657:2068241 [4] NCCL INFO comm 0xac13810 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb88431b88f65b221 - Init COMPLETE +ip-26-0-170-143:2067657:2068252 [4] NCCL INFO Channel 02/1 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2067657:2068252 [4] NCCL INFO Channel 03/1 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067655:2068239 [2] NCCL INFO comm 0x14e443b0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb88431b88f65b221 - Init COMPLETE +ip-26-0-170-143:2067655:2068253 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-143:2067655:2068253 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067656:2068240 [3] NCCL INFO comm 0x148e5840 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x464fd298d6760e0c - Init COMPLETE +ip-26-0-170-143:2067656:2068254 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2067656:2068254 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2067658:2068242 [5] NCCL INFO comm 0xbc1be20 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x464fd298d6760e0c - Init COMPLETE +ip-26-0-170-143:2067658:2068255 [5] NCCL INFO Channel 02/1 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-143:2067658:2068255 [5] NCCL INFO Channel 03/1 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO comm 0xb8f7b60 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8f37d53bcc14d66c - Init START +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO comm 0x13730490 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8f37d53bcc14d66c - Init START +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO comm 0xb83c0e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6e5d3e4511a94322 - Init START +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO comm 0x14904cc0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6e5d3e4511a94322 - Init START +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067659:2068266 [6] NCCL INFO comm 0xb8f7b60 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8f37d53bcc14d66c - Init COMPLETE +ip-26-0-170-143:2067659:2068277 [6] NCCL INFO Channel 02/1 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2067659:2068277 [6] NCCL INFO Channel 03/1 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067657:2068265 [4] NCCL INFO comm 0x13730490 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8f37d53bcc14d66c - Init COMPLETE +ip-26-0-170-143:2067657:2068278 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2067657:2068278 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2067660:2068267 [7] NCCL INFO comm 0xb83c0e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6e5d3e4511a94322 - Init COMPLETE +ip-26-0-170-143:2067660:2068279 [7] NCCL INFO Channel 02/1 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2067660:2068279 [7] NCCL INFO Channel 03/1 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2067658:2068264 [5] NCCL INFO comm 0x14904cc0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6e5d3e4511a94322 - Init COMPLETE +ip-26-0-170-143:2067658:2068280 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-143:2067658:2068280 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO comm 0xb9de540 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc54556f79e5cf38a - Init START +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO comm 0x14b17c90 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc54556f79e5cf38a - Init START +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO comm 0x141615b0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd44bcbadd5ad055e - Init START +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO comm 0xb275840 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd44bcbadd5ad055e - Init START +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Channel 00/0 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Channel 01/0 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Channel 02/0 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Channel 03/0 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 00/0 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 01/0 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 02/0 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 03/0 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 00/0 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 01/0 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 02/0 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 03/0 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Channel 00/0 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Channel 01/0 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Channel 02/0 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Channel 03/0 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067660:2068291 [7] NCCL INFO comm 0x14b17c90 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc54556f79e5cf38a - Init COMPLETE +ip-26-0-170-143:2067660:2068297 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2067660:2068297 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067659:2068290 [6] NCCL INFO comm 0x141615b0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd44bcbadd5ad055e - Init COMPLETE +ip-26-0-170-143:2067659:2068299 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2067659:2068299 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-160:2102339:2102907 [1] NCCL INFO comm 0xb9de540 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc54556f79e5cf38a - Init COMPLETE +ip-26-0-170-160:2102339:2102913 [1] NCCL INFO Channel 02/1 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-160:2102339:2102913 [1] NCCL INFO Channel 03/1 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-160:2102338:2102906 [0] NCCL INFO comm 0xb275840 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd44bcbadd5ad055e - Init COMPLETE +ip-26-0-170-160:2102338:2102914 [0] NCCL INFO Channel 02/1 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-160:2102338:2102914 [0] NCCL INFO Channel 03/1 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO comm 0x14376ed0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6ff12abab3e94732 - Init START +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO comm 0xc1e0150 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6ff12abab3e94732 - Init START +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO comm 0x13672820 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb503544ef508fd3 - Init START +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO comm 0xab36600 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb503544ef508fd3 - Init START +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102341:2102926 [3] NCCL INFO comm 0xc1e0150 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6ff12abab3e94732 - Init COMPLETE +ip-26-0-170-160:2102341:2102936 [3] NCCL INFO Channel 02/1 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-160:2102341:2102936 [3] NCCL INFO Channel 03/1 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-160:2102339:2102924 [1] NCCL INFO comm 0x14376ed0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6ff12abab3e94732 - Init COMPLETE +ip-26-0-170-160:2102339:2102937 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-160:2102339:2102937 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-160:2102340:2102925 [2] NCCL INFO comm 0xab36600 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb503544ef508fd3 - Init COMPLETE +ip-26-0-170-160:2102340:2102938 [2] NCCL INFO Channel 02/1 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-160:2102340:2102938 [2] NCCL INFO Channel 03/1 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-160:2102338:2102923 [0] NCCL INFO comm 0x13672820 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb503544ef508fd3 - Init COMPLETE +ip-26-0-170-160:2102338:2102939 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-160:2102338:2102939 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO comm 0xc5f0670 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbc511b101956e11c - Init START +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO comm 0x152a6210 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbc511b101956e11c - Init START +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO comm 0xc1c5070 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb6992407940b2d9c - Init START +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO comm 0x1304a0d0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb6992407940b2d9c - Init START +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Channel 00/0 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Channel 01/0 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Channel 02/0 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Channel 03/0 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Channel 00/0 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 00/0 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Channel 01/0 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Channel 02/0 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 01/0 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Channel 03/0 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 02/0 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 03/0 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 00/0 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 01/0 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 02/0 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 03/0 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102342:2102952 [4] NCCL INFO comm 0xc1c5070 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb6992407940b2d9c - Init COMPLETE +ip-26-0-170-160:2102342:2102962 [4] NCCL INFO Channel 02/1 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-160:2102342:2102962 [4] NCCL INFO Channel 03/1 : 0[2] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-160:2102343:2102951 [5] NCCL INFO comm 0xc5f0670 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbc511b101956e11c - Init COMPLETE +ip-26-0-170-160:2102343:2102963 [5] NCCL INFO Channel 02/1 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-160:2102343:2102963 [5] NCCL INFO Channel 03/1 : 0[3] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-160:2102341:2102949 [3] NCCL INFO comm 0x152a6210 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbc511b101956e11c - Init COMPLETE +ip-26-0-170-160:2102341:2102964 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-160:2102341:2102964 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[5] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-160:2102340:2102950 [2] NCCL INFO comm 0x1304a0d0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb6992407940b2d9c - Init COMPLETE +ip-26-0-170-160:2102340:2102965 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-160:2102340:2102965 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[4] [send] via NET/Libfabric/2/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO comm 0xcf30870 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcc1d6f3aae2f31ae - Init START +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO comm 0x1493cac0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcc1d6f3aae2f31ae - Init START +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO comm 0xc637d90 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf5324e03e67c2f4b - Init START +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO comm 0x15064790 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf5324e03e67c2f4b - Init START +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102344:2102975 [6] NCCL INFO comm 0xcf30870 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcc1d6f3aae2f31ae - Init COMPLETE +ip-26-0-170-160:2102344:2102987 [6] NCCL INFO Channel 02/1 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-160:2102344:2102987 [6] NCCL INFO Channel 03/1 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-160:2102342:2102974 [4] NCCL INFO comm 0x1493cac0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcc1d6f3aae2f31ae - Init COMPLETE +ip-26-0-170-160:2102342:2102988 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-160:2102342:2102988 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-160:2102345:2102976 [7] NCCL INFO comm 0xc637d90 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf5324e03e67c2f4b - Init COMPLETE +ip-26-0-170-160:2102345:2102989 [7] NCCL INFO Channel 02/1 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-160:2102345:2102989 [7] NCCL INFO Channel 03/1 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-160:2102343:2102973 [5] NCCL INFO comm 0x15064790 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf5324e03e67c2f4b - Init COMPLETE +ip-26-0-170-160:2102343:2102990 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-160:2102343:2102990 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-160:2102342:2103063 [4] NCCL INFO Channel 02/1 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-160:2102343:2103064 [5] NCCL INFO Channel 02/1 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-160:2102342:2103063 [4] NCCL INFO Channel 03/1 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-160:2102343:2103064 [5] NCCL INFO Channel 03/1 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:2102344:2103065 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-160:2102344:2103065 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-160:2102345:2103066 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-160:2102345:2103066 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-160:2102341:2103068 [3] NCCL INFO Channel 02/1 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-160:2102341:2103068 [3] NCCL INFO Channel 03/1 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-160:2102340:2103069 [2] NCCL INFO Channel 02/1 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-160:2102340:2103069 [2] NCCL INFO Channel 03/1 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-160:2102342:2103071 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-160:2102343:2103072 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-160:2102342:2103071 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-160:2102343:2103072 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-160:2102338:2103075 [0] NCCL INFO Channel 02/1 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-160:2102338:2103075 [0] NCCL INFO Channel 03/1 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-160:2102339:2103076 [1] NCCL INFO Channel 02/1 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-160:2102339:2103076 [1] NCCL INFO Channel 03/1 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-160:2102341:2103078 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-160:2102340:2103079 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-160:2102341:2103078 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-160:2102340:2103079 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-143:2067659:2068383 [6] NCCL INFO Channel 02/1 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2067660:2068384 [7] NCCL INFO Channel 02/1 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2067659:2068383 [6] NCCL INFO Channel 03/1 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2067660:2068384 [7] NCCL INFO Channel 03/1 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-160:2102338:2103081 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-160:2102339:2103082 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-160:2102338:2103081 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-160:2102339:2103082 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-143:2067658:2068386 [5] NCCL INFO Channel 02/1 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-143:2067658:2068386 [5] NCCL INFO Channel 03/1 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-143:2067657:2068387 [4] NCCL INFO Channel 02/1 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2067657:2068387 [4] NCCL INFO Channel 03/1 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2067659:2068388 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2067659:2068388 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2067660:2068389 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2067660:2068389 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2067656:2068391 [3] NCCL INFO Channel 02/1 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2067656:2068391 [3] NCCL INFO Channel 03/1 : 1[5] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2067655:2068392 [2] NCCL INFO Channel 02/1 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-143:2067655:2068392 [2] NCCL INFO Channel 03/1 : 1[4] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-143:2067658:2068394 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-143:2067658:2068394 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[3] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-143:2067657:2068395 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2067657:2068395 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[2] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2067653:2068398 [0] NCCL INFO Channel 02/1 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-143:2067653:2068398 [0] NCCL INFO Channel 03/1 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-143:2067654:2068399 [1] NCCL INFO Channel 02/1 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-143:2067654:2068399 [1] NCCL INFO Channel 03/1 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-143:2067656:2068400 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2067656:2068400 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2067655:2068401 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-143:2067655:2068401 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA/Shared +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +12/27/2024 15:54:38 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: Rank 7 P2P History: R1 R2 R3 R4 R5 R6 R7 R8 R9 S1 S2 S3 R10 R11 R12 R13 R14 R15 S4 S5 S6 R16 R17 R18 R19 R20 R21 S7 S8 S9 R22 R23 R24 R25 R26 R27 S10 S11 S12 R28 R29 R30 R31 R32 R33 S13 S14 S15 R34 R35 R36 R37 R38 R39 S16 S17 S18 R40 R41 R42 R43 R44 R45 S19 S20 S21 R46 R47 R48 R49 R50 R51 S22 S23 S24 R52 R53 R54 R55 R56 R57 S25 S26 S27 R58 R59 R60 R61 R62 R63 S28 S29 S30 R64 R65 R66 R67 R68 R69 S31 S32 S33 R70 R71 R72 R73 R74 R75 S34 S35 S36 R76 R77 R78 R79 R80 R81 S37 S38 S39 R82 R83 R84 R85 R86 R87 S40 S41 S42 R88 R89 R90 R91 R92 R93 S43 S44 S45 R94 R95 R96 S46 S47 S48 +12/27/2024 15:54:38 [INFO|DP=0|PP=7|TP=1|ip-26-0-170-160]: Rank 7 P2P History: R1 R2 R3 R4 R5 R6 R7 R8 R9 S1 S2 S3 R10 R11 R12 R13 R14 R15 S4 S5 S6 R16 R17 R18 R19 R20 R21 S7 S8 S9 R22 R23 R24 R25 R26 R27 S10 S11 S12 R28 R29 R30 R31 R32 R33 S13 S14 S15 R34 R35 R36 R37 R38 R39 S16 S17 S18 R40 R41 R42 R43 R44 R45 S19 S20 S21 R46 R47 R48 R49 R50 R51 S22 S23 S24 R52 R53 R54 R55 R56 R57 S25 S26 S27 R58 R59 R60 R61 R62 R63 S28 S29 S30 R64 R65 R66 R67 R68 R69 S31 S32 S33 R70 R71 R72 R73 R74 R75 S34 S35 S36 R76 R77 R78 R79 R80 R81 S37 S38 S39 R82 R83 R84 R85 R86 R87 S40 S41 S42 R88 R89 R90 R91 R92 R93 S43 S44 S45 R94 R95 R96 S46 S47 S48 +12/27/2024 15:54:38 [INFO|DP=0|PP=6|TP=0|ip-26-0-170-160]: Rank 6 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 R13 R14 R15 S10 S11 S12 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 R139 R140 R141 S136 S137 S138 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:38 [INFO|DP=0|PP=6|TP=1|ip-26-0-170-160]: Rank 6 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 R13 R14 R15 S10 S11 S12 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 R139 R140 R141 S136 S137 S138 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:38 [INFO|DP=0|PP=5|TP=0|ip-26-0-170-160]: Rank 5 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:38 [INFO|DP=0|PP=5|TP=1|ip-26-0-170-160]: Rank 5 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:38 [INFO|DP=0|PP=4|TP=1|ip-26-0-170-160]: Rank 4 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:38 [INFO|DP=0|PP=4|TP=0|ip-26-0-170-160]: Rank 4 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:38 [INFO|DP=0|PP=3|TP=0|ip-26-0-170-143]: Rank 3 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:38 [INFO|DP=0|PP=3|TP=1|ip-26-0-170-143]: Rank 3 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:39 [INFO|DP=0|PP=2|TP=0|ip-26-0-170-143]: Rank 2 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:39 [INFO|DP=0|PP=2|TP=1|ip-26-0-170-143]: Rank 2 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:39 [INFO|DP=0|PP=1|TP=1|ip-26-0-170-143]: Rank 1 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 S34 S35 S36 R37 R38 R39 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 S124 S125 S126 R127 R128 R129 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:39 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-143]: Rank 1 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 S34 S35 S36 R37 R38 R39 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 S124 S125 S126 R127 R128 R129 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:54:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Rank 0 P2P History: S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 S19 S20 S21 S22 S23 S24 S25 S26 S27 S28 S29 S30 S31 S32 S33 S34 S35 S36 S37 S38 S39 S40 S41 S42 S43 S44 S45 R1 R2 R3 S46 S47 S48 S49 S50 S51 R4 R5 R6 S52 S53 S54 S55 S56 S57 R7 R8 R9 S58 S59 S60 S61 S62 S63 R10 R11 R12 S64 S65 S66 S67 S68 S69 R13 R14 R15 S70 S71 S72 S73 S74 S75 R16 R17 R18 S76 S77 S78 S79 S80 S81 R19 R20 R21 S82 S83 S84 S85 S86 S87 R22 R23 R24 S88 S89 S90 S91 S92 S93 R25 R26 R27 S94 S95 S96 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 R40 R41 R42 R43 R44 R45 R46 R47 R48 +12/27/2024 15:54:39 [INFO|DP=0|PP=0|TP=1|ip-26-0-170-143]: Rank 0 P2P History: S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 S19 S20 S21 S22 S23 S24 S25 S26 S27 S28 S29 S30 S31 S32 S33 S34 S35 S36 S37 S38 S39 S40 S41 S42 S43 S44 S45 R1 R2 R3 S46 S47 S48 S49 S50 S51 R4 R5 R6 S52 S53 S54 S55 S56 S57 R7 R8 R9 S58 S59 S60 S61 S62 S63 R10 R11 R12 S64 S65 S66 S67 S68 S69 R13 R14 R15 S70 S71 S72 S73 S74 S75 R16 R17 R18 S76 S77 S78 S79 S80 S81 R19 R20 R21 S82 S83 S84 S85 S86 S87 R22 R23 R24 S88 S89 S90 S91 S92 S93 R25 R26 R27 S94 S95 S96 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 R40 R41 R42 R43 R44 R45 R46 R47 R48 +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO comm 0x1ffc6da0 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO comm 0x222892e0 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO comm 0x16618f70 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO comm 0x15ef36c0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO comm 0x1687b110 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO comm 0x1461e740 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO comm 0x14c817d0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO comm 0x15983180 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +12/27/2024 15:54:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 1728.62MiB. Peak allocated 5221.45MiB. Peak reserved: 5528.00MiB +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO comm 0x1cc21d90 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO comm 0x15f25ac0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO comm 0x15763aa0 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO comm 0x14d4e950 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO comm 0x164d5d30 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO comm 0x1611a710 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO comm 0x15f774d0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO comm 0x1dee18a0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x40e22b9775e75fd5 - Init START +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 7/-1/-1->15->-1 [3] 7/-1/-1->15->-1 +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Trees [0] 8/-1/-1->0->-1 [1] 8/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Trees [0] 4/12/-1->8->0 [1] 4/12/-1->8->0 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102345:2103109 [7] NCCL INFO comm 0x1ffc6da0 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067653:2068423 [0] NCCL INFO comm 0x1dee18a0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102344:2103108 [6] NCCL INFO comm 0x222892e0 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102340:2103112 [2] NCCL INFO comm 0x1461e740 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102343:2103110 [5] NCCL INFO comm 0x16618f70 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102342:2103111 [4] NCCL INFO comm 0x15ef36c0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-160:2102339:2103114 [1] NCCL INFO comm 0x15983180 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-160:2102338:2103113 [0] NCCL INFO comm 0x14c817d0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067654:2068424 [1] NCCL INFO comm 0x1cc21d90 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067659:2068427 [6] NCCL INFO comm 0x15763aa0 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067655:2068428 [2] NCCL INFO comm 0x164d5d30 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067656:2068430 [3] NCCL INFO comm 0x15f774d0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-160:2102341:2103115 [3] NCCL INFO comm 0x1687b110 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2067660:2068426 [7] NCCL INFO comm 0x1611a710 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-143:2067658:2068425 [5] NCCL INFO comm 0x15f25ac0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x40e22b9775e75fd5 - Init COMPLETE +ip-26-0-170-143:2067657:2068429 [4] NCCL INFO comm 0x14d4e950 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x40e22b9775e75fd5 - Init COMPLETE +12/27/2024 15:54:45 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 3008.65MiB. Peak allocated 3008.66MiB. Peak reserved: 6552.00MiB +12/27/2024 15:54:45 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: iteration: 1 / 100 | consumed_tokens: 131K | elapsed_time_per_iteration_ms: 46.8K | tokens_per_sec: 2.8K | tokens_per_sec_per_gpu: 175 | global_batch_size: 32 | lm_loss: 12 | lr: 0.00015 | model_tflops_per_gpu: 1.69 | hardware_tflops_per_gpu: 1.69 | grad_norm: 1.89 | cuda_memory_allocated: 3.15G | cuda_max_memory_reserved: 9.82G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.9G | hd_free_memory_tb: 242G +12/27/2024 15:55:04 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: Rank 7 P2P History: R1 R2 R3 R4 R5 R6 R7 R8 R9 S1 S2 S3 R10 R11 R12 R13 R14 R15 S4 S5 S6 R16 R17 R18 R19 R20 R21 S7 S8 S9 R22 R23 R24 R25 R26 R27 S10 S11 S12 R28 R29 R30 R31 R32 R33 S13 S14 S15 R34 R35 R36 R37 R38 R39 S16 S17 S18 R40 R41 R42 R43 R44 R45 S19 S20 S21 R46 R47 R48 R49 R50 R51 S22 S23 S24 R52 R53 R54 R55 R56 R57 S25 S26 S27 R58 R59 R60 R61 R62 R63 S28 S29 S30 R64 R65 R66 R67 R68 R69 S31 S32 S33 R70 R71 R72 R73 R74 R75 S34 S35 S36 R76 R77 R78 R79 R80 R81 S37 S38 S39 R82 R83 R84 R85 R86 R87 S40 S41 S42 R88 R89 R90 R91 R92 R93 S43 S44 S45 R94 R95 R96 S46 S47 S48 +12/27/2024 15:55:04 [INFO|DP=0|PP=7|TP=1|ip-26-0-170-160]: Rank 7 P2P History: R1 R2 R3 R4 R5 R6 R7 R8 R9 S1 S2 S3 R10 R11 R12 R13 R14 R15 S4 S5 S6 R16 R17 R18 R19 R20 R21 S7 S8 S9 R22 R23 R24 R25 R26 R27 S10 S11 S12 R28 R29 R30 R31 R32 R33 S13 S14 S15 R34 R35 R36 R37 R38 R39 S16 S17 S18 R40 R41 R42 R43 R44 R45 S19 S20 S21 R46 R47 R48 R49 R50 R51 S22 S23 S24 R52 R53 R54 R55 R56 R57 S25 S26 S27 R58 R59 R60 R61 R62 R63 S28 S29 S30 R64 R65 R66 R67 R68 R69 S31 S32 S33 R70 R71 R72 R73 R74 R75 S34 S35 S36 R76 R77 R78 R79 R80 R81 S37 S38 S39 R82 R83 R84 R85 R86 R87 S40 S41 S42 R88 R89 R90 R91 R92 R93 S43 S44 S45 R94 R95 R96 S46 S47 S48 +12/27/2024 15:55:04 [INFO|DP=0|PP=6|TP=0|ip-26-0-170-160]: Rank 6 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 R13 R14 R15 S10 S11 S12 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 R139 R140 R141 S136 S137 S138 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:04 [INFO|DP=0|PP=6|TP=1|ip-26-0-170-160]: Rank 6 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 R13 R14 R15 S10 S11 S12 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 R139 R140 R141 S136 S137 S138 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:04 [INFO|DP=0|PP=5|TP=1|ip-26-0-170-160]: Rank 5 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:04 [INFO|DP=0|PP=5|TP=0|ip-26-0-170-160]: Rank 5 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:05 [INFO|DP=0|PP=4|TP=0|ip-26-0-170-160]: Rank 4 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:05 [INFO|DP=0|PP=4|TP=1|ip-26-0-170-160]: Rank 4 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:05 [INFO|DP=0|PP=3|TP=0|ip-26-0-170-143]: Rank 3 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:05 [INFO|DP=0|PP=3|TP=1|ip-26-0-170-143]: Rank 3 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:05 [INFO|DP=0|PP=2|TP=1|ip-26-0-170-143]: Rank 2 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:05 [INFO|DP=0|PP=2|TP=0|ip-26-0-170-143]: Rank 2 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:05 [INFO|DP=0|PP=1|TP=1|ip-26-0-170-143]: Rank 1 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 S34 S35 S36 R37 R38 R39 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 S124 S125 S126 R127 R128 R129 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:05 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-143]: Rank 1 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 S34 S35 S36 R37 R38 R39 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 S124 S125 S126 R127 R128 R129 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:05 [INFO|DP=0|PP=0|TP=1|ip-26-0-170-143]: Rank 0 P2P History: S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 S19 S20 S21 S22 S23 S24 S25 S26 S27 S28 S29 S30 S31 S32 S33 S34 S35 S36 S37 S38 S39 S40 S41 S42 S43 S44 S45 R1 R2 R3 S46 S47 S48 S49 S50 S51 R4 R5 R6 S52 S53 S54 S55 S56 S57 R7 R8 R9 S58 S59 S60 S61 S62 S63 R10 R11 R12 S64 S65 S66 S67 S68 S69 R13 R14 R15 S70 S71 S72 S73 S74 S75 R16 R17 R18 S76 S77 S78 S79 S80 S81 R19 R20 R21 S82 S83 S84 S85 S86 S87 R22 R23 R24 S88 S89 S90 S91 S92 S93 R25 R26 R27 S94 S95 S96 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 R40 R41 R42 R43 R44 R45 R46 R47 R48 +12/27/2024 15:55:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Rank 0 P2P History: S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 S19 S20 S21 S22 S23 S24 S25 S26 S27 S28 S29 S30 S31 S32 S33 S34 S35 S36 S37 S38 S39 S40 S41 S42 S43 S44 S45 R1 R2 R3 S46 S47 S48 S49 S50 S51 R4 R5 R6 S52 S53 S54 S55 S56 S57 R7 R8 R9 S58 S59 S60 S61 S62 S63 R10 R11 R12 S64 S65 S66 S67 S68 S69 R13 R14 R15 S70 S71 S72 S73 S74 S75 R16 R17 R18 S76 S77 S78 S79 S80 S81 R19 R20 R21 S82 S83 S84 S85 S86 S87 R22 R23 R24 S88 S89 S90 S91 S92 S93 R25 R26 R27 S94 S95 S96 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 R40 R41 R42 R43 R44 R45 R46 R47 R48 +12/27/2024 15:55:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 3008.65MiB. Peak allocated 6501.49MiB. Peak reserved: 7128.00MiB +12/27/2024 15:55:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 3008.65MiB. Peak allocated 3008.66MiB. Peak reserved: 7128.00MiB +12/27/2024 15:55:05 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: iteration: 2 / 100 | consumed_tokens: 262K | elapsed_time_per_iteration_ms: 20.2K | tokens_per_sec: 6.49K | tokens_per_sec_per_gpu: 406 | global_batch_size: 32 | lm_loss: 12 | lr: 0.0003 | model_tflops_per_gpu: 3.92 | hardware_tflops_per_gpu: 3.92 | grad_norm: 1.9 | cuda_memory_allocated: 3.15G | cuda_max_memory_reserved: 9.82G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.9G | hd_free_memory_tb: 242G +12/27/2024 15:55:23 [INFO|DP=0|PP=7|TP=1|ip-26-0-170-160]: Rank 7 P2P History: R1 R2 R3 R4 R5 R6 R7 R8 R9 S1 S2 S3 R10 R11 R12 R13 R14 R15 S4 S5 S6 R16 R17 R18 R19 R20 R21 S7 S8 S9 R22 R23 R24 R25 R26 R27 S10 S11 S12 R28 R29 R30 R31 R32 R33 S13 S14 S15 R34 R35 R36 R37 R38 R39 S16 S17 S18 R40 R41 R42 R43 R44 R45 S19 S20 S21 R46 R47 R48 R49 R50 R51 S22 S23 S24 R52 R53 R54 R55 R56 R57 S25 S26 S27 R58 R59 R60 R61 R62 R63 S28 S29 S30 R64 R65 R66 R67 R68 R69 S31 S32 S33 R70 R71 R72 R73 R74 R75 S34 S35 S36 R76 R77 R78 R79 R80 R81 S37 S38 S39 R82 R83 R84 R85 R86 R87 S40 S41 S42 R88 R89 R90 R91 R92 R93 S43 S44 S45 R94 R95 R96 S46 S47 S48 +12/27/2024 15:55:23 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: Rank 7 P2P History: R1 R2 R3 R4 R5 R6 R7 R8 R9 S1 S2 S3 R10 R11 R12 R13 R14 R15 S4 S5 S6 R16 R17 R18 R19 R20 R21 S7 S8 S9 R22 R23 R24 R25 R26 R27 S10 S11 S12 R28 R29 R30 R31 R32 R33 S13 S14 S15 R34 R35 R36 R37 R38 R39 S16 S17 S18 R40 R41 R42 R43 R44 R45 S19 S20 S21 R46 R47 R48 R49 R50 R51 S22 S23 S24 R52 R53 R54 R55 R56 R57 S25 S26 S27 R58 R59 R60 R61 R62 R63 S28 S29 S30 R64 R65 R66 R67 R68 R69 S31 S32 S33 R70 R71 R72 R73 R74 R75 S34 S35 S36 R76 R77 R78 R79 R80 R81 S37 S38 S39 R82 R83 R84 R85 R86 R87 S40 S41 S42 R88 R89 R90 R91 R92 R93 S43 S44 S45 R94 R95 R96 S46 S47 S48 +12/27/2024 15:55:24 [INFO|DP=0|PP=6|TP=1|ip-26-0-170-160]: Rank 6 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 R13 R14 R15 S10 S11 S12 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 R139 R140 R141 S136 S137 S138 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:24 [INFO|DP=0|PP=6|TP=0|ip-26-0-170-160]: Rank 6 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 R13 R14 R15 S10 S11 S12 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 R139 R140 R141 S136 S137 S138 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:24 [INFO|DP=0|PP=5|TP=1|ip-26-0-170-160]: Rank 5 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:24 [INFO|DP=0|PP=5|TP=0|ip-26-0-170-160]: Rank 5 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 R19 R20 R21 S16 S17 S18 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 R136 R137 R138 S133 S134 S135 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:24 [INFO|DP=0|PP=4|TP=1|ip-26-0-170-160]: Rank 4 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:24 [INFO|DP=0|PP=4|TP=0|ip-26-0-170-160]: Rank 4 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 R25 R26 R27 S22 S23 S24 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 R133 R134 R135 S130 S131 S132 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:24 [INFO|DP=0|PP=3|TP=1|ip-26-0-170-143]: Rank 3 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:24 [INFO|DP=0|PP=3|TP=0|ip-26-0-170-143]: Rank 3 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 R31 R32 R33 S28 S29 S30 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 R130 R131 R132 S127 S128 S129 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:24 [INFO|DP=0|PP=2|TP=1|ip-26-0-170-143]: Rank 2 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:24 [INFO|DP=0|PP=2|TP=0|ip-26-0-170-143]: Rank 2 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 R37 R38 R39 S34 S35 S36 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 R127 R128 R129 S124 S125 S126 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:24 [INFO|DP=0|PP=1|TP=1|ip-26-0-170-143]: Rank 1 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 S34 S35 S36 R37 R38 R39 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 S124 S125 S126 R127 R128 R129 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:24 [INFO|DP=0|PP=0|TP=1|ip-26-0-170-143]: Rank 0 P2P History: S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 S19 S20 S21 S22 S23 S24 S25 S26 S27 S28 S29 S30 S31 S32 S33 S34 S35 S36 S37 S38 S39 S40 S41 S42 S43 S44 S45 R1 R2 R3 S46 S47 S48 S49 S50 S51 R4 R5 R6 S52 S53 S54 S55 S56 S57 R7 R8 R9 S58 S59 S60 S61 S62 S63 R10 R11 R12 S64 S65 S66 S67 S68 S69 R13 R14 R15 S70 S71 S72 S73 S74 S75 R16 R17 R18 S76 S77 S78 S79 S80 S81 R19 R20 R21 S82 S83 S84 S85 S86 S87 R22 R23 R24 S88 S89 S90 S91 S92 S93 R25 R26 R27 S94 S95 S96 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 R40 R41 R42 R43 R44 R45 R46 R47 R48 +12/27/2024 15:55:24 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Rank 0 P2P History: S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 S19 S20 S21 S22 S23 S24 S25 S26 S27 S28 S29 S30 S31 S32 S33 S34 S35 S36 S37 S38 S39 S40 S41 S42 S43 S44 S45 R1 R2 R3 S46 S47 S48 S49 S50 S51 R4 R5 R6 S52 S53 S54 S55 S56 S57 R7 R8 R9 S58 S59 S60 S61 S62 S63 R10 R11 R12 S64 S65 S66 S67 S68 S69 R13 R14 R15 S70 S71 S72 S73 S74 S75 R16 R17 R18 S76 S77 S78 S79 S80 S81 R19 R20 R21 S82 S83 S84 S85 S86 S87 R22 R23 R24 S88 S89 S90 S91 S92 S93 R25 R26 R27 S94 S95 S96 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 R40 R41 R42 R43 R44 R45 R46 R47 R48 +12/27/2024 15:55:24 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 3008.65MiB. Peak allocated 6501.49MiB. Peak reserved: 7128.00MiB +12/27/2024 15:55:25 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-143]: Rank 1 P2P History: R1 R2 R3 R4 R5 R6 S1 S2 S3 S4 S5 S6 R7 R8 R9 R10 R11 R12 S7 S8 S9 S10 S11 S12 R13 R14 R15 R16 R17 R18 S13 S14 S15 S16 S17 S18 R19 R20 R21 R22 R23 R24 S19 S20 S21 S22 S23 S24 R25 R26 R27 R28 R29 R30 S25 S26 S27 S28 S29 S30 R31 R32 R33 R34 R35 R36 S31 S32 S33 S34 S35 S36 R37 R38 R39 R40 R41 R42 S37 S38 S39 R43 R44 R45 S40 S41 S42 R46 R47 R48 S43 S44 S45 R49 R50 R51 S46 S47 S48 R52 R53 R54 S49 S50 S51 R55 R56 R57 S52 S53 S54 R58 R59 R60 S55 S56 S57 R61 R62 R63 S58 S59 S60 R64 R65 R66 S61 S62 S63 R67 R68 R69 S64 S65 S66 R70 R71 R72 S67 S68 S69 R73 R74 R75 S70 S71 S72 R76 R77 R78 S73 S74 S75 R79 R80 R81 S76 S77 S78 R82 R83 R84 S79 S80 S81 R85 R86 R87 S82 S83 S84 R88 R89 R90 S85 S86 S87 R91 R92 R93 S88 S89 S90 R94 R95 R96 S91 S92 S93 R97 R98 R99 S94 S95 S96 R100 R101 R102 S97 S98 S99 R103 R104 R105 S100 S101 S102 R106 R107 R108 S103 S104 S105 R109 R110 R111 S106 S107 S108 R112 R113 R114 S109 S110 S111 R115 R116 R117 S112 S113 S114 R118 R119 R120 S115 S116 S117 R121 R122 R123 S118 S119 S120 R124 R125 R126 S121 S122 S123 S124 S125 S126 R127 R128 R129 S127 S128 S129 R130 R131 R132 S130 S131 S132 R133 R134 R135 S133 S134 S135 R136 R137 R138 S136 S137 S138 R139 R140 R141 S139 S140 S141 R142 R143 R144 S142 S143 S144 +12/27/2024 15:55:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/27/2024 15:55:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: | -------- | -------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | ----------------------------------- | -------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/27/2024 15:55:25 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: iteration: 3 / 100 | consumed_tokens: 393K | elapsed_time_per_iteration_ms: 19.3K | tokens_per_sec: 6.78K | tokens_per_sec_per_gpu: 424 | global_batch_size: 32 | lm_loss: 12 | lr: 0.000296 | model_tflops_per_gpu: 4.1 | hardware_tflops_per_gpu: 4.1 | grad_norm: 1.88 | cuda_memory_allocated: 3.15G | cuda_max_memory_reserved: 9.82G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.9G | hd_free_memory_tb: 242G +12/27/2024 15:55:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: | 13767063 | 1.34G_dp1_tp2_pp8_acc16_mbs2_seq4096_zero1_tpmodeRED_vocab131k | 2 | 4096 | 2 | 16 | 32 | 4.10 | 4.10 | 423.81 | 45.43 | 41.86 | 42.28 | 45.35 | 41.91 | 41.14 | 2.94 | 6.96 | 1 | 8 | 2 | OneForwardOneBackwardPipelineEngine | REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 32 | 4096 | 131072 | True | torch.bfloat16 | 1 | 25 | True | 1.61G | 168M | +12/27/2024 15:55:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final2_pp.csv +12/27/2024 15:55:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Acquired lock for CSV file: benchmark/results/bench_final2_pp.csv +12/27/2024 15:55:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Successfully wrote to CSV file: benchmark/results/bench_final2_pp.csv. Releasing lock... +12/27/2024 15:55:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +[2024-12-27 15:55:55,816] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2102338) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 15:55:55,824] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2067653) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-160.ec2.internal + rank : 9 (local_rank: 1) + exitcode : 1 (pid: 2102339) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-160.ec2.internal + rank : 10 (local_rank: 2) + exitcode : 1 (pid: 2102340) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-160.ec2.internal + rank : 11 (local_rank: 3) + exitcode : 1 (pid: 2102341) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-160.ec2.internal + rank : 12 (local_rank: 4) + exitcode : 1 (pid: 2102342) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-160.ec2.internal + rank : 13 (local_rank: 5) + exitcode : 1 (pid: 2102343) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-160.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 2102344) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-160.ec2.internal + rank : 15 (local_rank: 7) + exitcode : 1 (pid: 2102345) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-160.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 2102338) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-143.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 2067654) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-143.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 2067655) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-143.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 2067656) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-143.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 2067657) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-143.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 2067658) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-143.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 2067659) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-143.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 2067660) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_15:55:55 + host : ip-26-0-170-143.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 2067653) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-170-160: task 1: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13767063.0 +srun: error: ip-26-0-170-143: task 0: Exited with exit code 1 diff --git a/logs/13803153-bench_8.86G_dp2_tp2_pp2_acc2_mbs64_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13803153-bench_8.86G_dp2_tp2_pp2_acc2_mbs64_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..ee5a503cbad0e05b938ba55e3673221982b2eaf9 --- /dev/null +++ b/logs/13803153-bench_8.86G_dp2_tp2_pp2_acc2_mbs64_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,2710 @@ ++ '[' -z 13803153 ']' ++ unset FI_PROVIDER ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames ip-26-0-172-252 ++ export NODELIST=ip-26-0-172-252 ++ NODELIST=ip-26-0-172-252 +++ scontrol show hostnames ip-26-0-172-252 +++ head -n1 ++ export MASTER_NODE=ip-26-0-172-252 ++ MASTER_NODE=ip-26-0-172-252 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=1 ++ NNODES=1 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=8 ++ WORLD_SIZE=8 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ export NCCL_P2P_LEVEL=LOC ++ NCCL_P2P_LEVEL=LOC ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-172-252' +Master node: ip-26-0-172-252 ++ echo 'All nodes: ip-26-0-172-252' +All nodes: ip-26-0-172-252 ++ echo 'World size: 8' +World size: 8 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=1 --nproc_per_node=8 --rdzv_id=13803153 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-172-252:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_8.86G_dp2_tp2_pp2_acc2_mbs64_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-27 17:22:51,948] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:22:51,948] torch.distributed.run: [WARNING] +[2024-12-27 17:22:51,948] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:22:51,948] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:22:51,948] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-172-252:1108193:1108193 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.172.252<0> +ip-26-0-172-252:1108193:1108193 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-252:1108193:1108193 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-252:1108193:1108193 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-252:1108195:1108195 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-252:1108199:1108199 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-252:1108198:1108198 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-252:1108197:1108197 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-252:1108196:1108196 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-252:1108194:1108194 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-252:1108195:1108195 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.172.252<0> +ip-26-0-172-252:1108200:1108200 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-252:1108195:1108195 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-252:1108195:1108195 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-252:1108198:1108198 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.172.252<0> +ip-26-0-172-252:1108198:1108198 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-252:1108198:1108198 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-252:1108197:1108197 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.172.252<0> +ip-26-0-172-252:1108199:1108199 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.172.252<0> +ip-26-0-172-252:1108194:1108194 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.172.252<0> +ip-26-0-172-252:1108196:1108196 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.172.252<0> +ip-26-0-172-252:1108197:1108197 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-252:1108197:1108197 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-252:1108200:1108200 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.172.252<0> +ip-26-0-172-252:1108196:1108196 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-252:1108194:1108194 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-252:1108199:1108199 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-252:1108196:1108196 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-252:1108199:1108199 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-252:1108194:1108194 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-252:1108200:1108200 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-252:1108200:1108200 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/168 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/168 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/168 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/181 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/168 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/168 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/168 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/168 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO comm 0xa0bfb20 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x446a2e7fa6afac84 - Init START +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO comm 0x8c7f810 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x446a2e7fa6afac84 - Init START +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO comm 0x94f84e0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x446a2e7fa6afac84 - Init START +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO comm 0x9a07e10 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x446a2e7fa6afac84 - Init START +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO comm 0x9c8f440 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x446a2e7fa6afac84 - Init START +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO comm 0x9bf0030 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x446a2e7fa6afac84 - Init START +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO comm 0xa2cb2f0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x446a2e7fa6afac84 - Init START +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO comm 0x8734f60 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x446a2e7fa6afac84 - Init START +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/181 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Connected all rings +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Connected all rings +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Connected all rings +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Connected all rings +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Connected all rings +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Connected all rings +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Connected all rings +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Connected all rings +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO Connected all trees +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO Connected all trees +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108200:1108363 [7] NCCL INFO comm 0x94f84e0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x446a2e7fa6afac84 - Init COMPLETE +ip-26-0-172-252:1108193:1108360 [0] NCCL INFO comm 0x9a07e10 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x446a2e7fa6afac84 - Init COMPLETE +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO Connected all trees +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO Connected all trees +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108195:1108361 [2] NCCL INFO comm 0xa2cb2f0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x446a2e7fa6afac84 - Init COMPLETE +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO Connected all trees +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108199:1108366 [6] NCCL INFO comm 0x9bf0030 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x446a2e7fa6afac84 - Init COMPLETE +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO Connected all trees +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108194:1108362 [1] NCCL INFO comm 0x8734f60 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x446a2e7fa6afac84 - Init COMPLETE +ip-26-0-172-252:1108198:1108367 [5] NCCL INFO comm 0xa0bfb20 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x446a2e7fa6afac84 - Init COMPLETE +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO Connected all trees +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO Connected all trees +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108196:1108365 [3] NCCL INFO comm 0x9c8f440 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x446a2e7fa6afac84 - Init COMPLETE +ip-26-0-172-252:1108197:1108364 [4] NCCL INFO comm 0x8c7f810 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x446a2e7fa6afac84 - Init COMPLETE +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO comm 0x8eb27e0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x14b82f8f4babbf86 - Init START +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO comm 0xa2f2b80 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x14b82f8f4babbf86 - Init START +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO comm 0x9ec2500 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x14b82f8f4babbf86 - Init START +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO comm 0x8968700 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x14b82f8f4babbf86 - Init START +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO comm 0x9c3c310 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x14b82f8f4babbf86 - Init START +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO comm 0x9e23180 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x14b82f8f4babbf86 - Init START +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO comm 0xa4fd640 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x14b82f8f4babbf86 - Init START +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO comm 0x972af60 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x14b82f8f4babbf86 - Init START +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/181 +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Connected all rings +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Connected all rings +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Connected all rings +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Connected all rings +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Connected all rings +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Connected all rings +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Connected all rings +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Connected all rings +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO Connected all trees +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108200:1108593 [7] NCCL INFO comm 0x972af60 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x14b82f8f4babbf86 - Init COMPLETE +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO Connected all trees +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108193:1108590 [0] NCCL INFO comm 0x9c3c310 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x14b82f8f4babbf86 - Init COMPLETE +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO Connected all trees +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108194:1108597 [1] NCCL INFO comm 0x8968700 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x14b82f8f4babbf86 - Init COMPLETE +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO Connected all trees +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO Connected all trees +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108199:1108595 [6] NCCL INFO comm 0x9e23180 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x14b82f8f4babbf86 - Init COMPLETE +ip-26-0-172-252:1108195:1108596 [2] NCCL INFO comm 0xa4fd640 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x14b82f8f4babbf86 - Init COMPLETE +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO Connected all trees +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO Connected all trees +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO Connected all trees +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108196:1108594 [3] NCCL INFO comm 0x9ec2500 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x14b82f8f4babbf86 - Init COMPLETE +ip-26-0-172-252:1108198:1108592 [5] NCCL INFO comm 0xa2f2b80 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x14b82f8f4babbf86 - Init COMPLETE +ip-26-0-172-252:1108197:1108591 [4] NCCL INFO comm 0x8eb27e0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x14b82f8f4babbf86 - Init COMPLETE +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO comm 0xa307fa0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2c1fb501e08f200f - Init START +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO comm 0xa512a80 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2c1fb501e08f200f - Init START +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO comm 0x9e38600 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2c1fb501e08f200f - Init START +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO comm 0x9ed7a00 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2c1fb501e08f200f - Init START +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO comm 0x8ec7940 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2c1fb501e08f200f - Init START +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO comm 0x973fa60 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2c1fb501e08f200f - Init START +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO comm 0x9c52680 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2c1fb501e08f200f - Init START +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO comm 0x897da00 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2c1fb501e08f200f - Init START +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/181 +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Connected all rings +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Connected all rings +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Connected all rings +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Connected all rings +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Connected all rings +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Connected all rings +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Connected all rings +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Connected all rings +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO Connected all trees +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108200:1108627 [7] NCCL INFO comm 0x973fa60 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2c1fb501e08f200f - Init COMPLETE +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO Connected all trees +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108193:1108624 [0] NCCL INFO comm 0x9c52680 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2c1fb501e08f200f - Init COMPLETE +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO Connected all trees +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO Connected all trees +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108199:1108629 [6] NCCL INFO comm 0x9e38600 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2c1fb501e08f200f - Init COMPLETE +ip-26-0-172-252:1108194:1108631 [1] NCCL INFO comm 0x897da00 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2c1fb501e08f200f - Init COMPLETE +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO Connected all trees +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO Connected all trees +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO Connected all trees +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO Connected all trees +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108196:1108630 [3] NCCL INFO comm 0x9ed7a00 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2c1fb501e08f200f - Init COMPLETE +ip-26-0-172-252:1108195:1108626 [2] NCCL INFO comm 0xa512a80 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2c1fb501e08f200f - Init COMPLETE +ip-26-0-172-252:1108198:1108625 [5] NCCL INFO comm 0xa307fa0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2c1fb501e08f200f - Init COMPLETE +ip-26-0-172-252:1108197:1108628 [4] NCCL INFO comm 0x8ec7940 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2c1fb501e08f200f - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO comm 0xbf8ef40 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8b2cd93628313dd8 - Init START +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO comm 0xab4cfa0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8b2cd93628313dd8 - Init START +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO comm 0xb3c3ae0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd2fc4aa6bf0b486f - Init START +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO comm 0xbabc590 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd2fc4aa6bf0b486f - Init START +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO comm 0xa7cc020 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb3630a9a629dfea - Init START +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO comm 0xbaa3360 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb3630a9a629dfea - Init START +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/181 +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO comm 0xc361b90 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8ad2b6626b5c6a95 - Init START +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO comm 0xbd28d90 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8ad2b6626b5c6a95 - Init START +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Connected all rings +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO Connected all trees +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Connected all rings +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO Connected all trees +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Connected all rings +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO Connected all trees +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Connected all rings +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO Connected all trees +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Connected all rings +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO Connected all trees +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Connected all rings +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO Connected all trees +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108199:1108665 [6] NCCL INFO comm 0xbabc590 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd2fc4aa6bf0b486f - Init COMPLETE +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Connected all rings +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO Connected all trees +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Connected all rings +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO Connected all trees +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108200:1108668 [7] NCCL INFO comm 0xb3c3ae0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd2fc4aa6bf0b486f - Init COMPLETE +ip-26-0-172-252:1108193:1108664 [0] NCCL INFO comm 0xbaa3360 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb3630a9a629dfea - Init COMPLETE +ip-26-0-172-252:1108194:1108669 [1] NCCL INFO comm 0xa7cc020 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb3630a9a629dfea - Init COMPLETE +ip-26-0-172-252:1108197:1108671 [4] NCCL INFO comm 0xab4cfa0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8b2cd93628313dd8 - Init COMPLETE +ip-26-0-172-252:1108195:1108666 [2] NCCL INFO comm 0xc361b90 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8ad2b6626b5c6a95 - Init COMPLETE +ip-26-0-172-252:1108196:1108670 [3] NCCL INFO comm 0xbd28d90 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8ad2b6626b5c6a95 - Init COMPLETE +ip-26-0-172-252:1108198:1108672 [5] NCCL INFO comm 0xbf8ef40 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8b2cd93628313dd8 - Init COMPLETE +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO comm 0xbab6300 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x721e6042a5cc8ff7 - Init START +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO comm 0xab5ff40 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x721e6042a5cc8ff7 - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/181 +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO comm 0xbacfc70 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9eeb6c25104bc986 - Init START +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO comm 0xc374b30 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9eeb6c25104bc986 - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO comm 0xbd3bc90 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8298dc332eb6c408 - Init START +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO comm 0xb3d69e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8298dc332eb6c408 - Init START +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO comm 0xbfa1e40 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc87a836ade5bb617 - Init START +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO comm 0xa7df220 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc87a836ade5bb617 - Init START +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Connected all rings +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO Connected all trees +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Connected all rings +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO Connected all trees +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Connected all rings +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO Connected all trees +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108199:1108696 [6] NCCL INFO comm 0xbacfc70 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9eeb6c25104bc986 - Init COMPLETE +ip-26-0-172-252:1108200:1108699 [7] NCCL INFO comm 0xb3d69e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8298dc332eb6c408 - Init COMPLETE +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Connected all rings +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO Connected all trees +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Connected all rings +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO Connected all trees +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108195:1108695 [2] NCCL INFO comm 0xc374b30 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9eeb6c25104bc986 - Init COMPLETE +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Connected all rings +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO Connected all trees +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108194:1108701 [1] NCCL INFO comm 0xa7df220 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc87a836ade5bb617 - Init COMPLETE +ip-26-0-172-252:1108196:1108698 [3] NCCL INFO comm 0xbd3bc90 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8298dc332eb6c408 - Init COMPLETE +ip-26-0-172-252:1108198:1108702 [5] NCCL INFO comm 0xbfa1e40 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc87a836ade5bb617 - Init COMPLETE +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Connected all rings +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO Connected all trees +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Connected all rings +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO Connected all trees +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108197:1108693 [4] NCCL INFO comm 0xab5ff40 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x721e6042a5cc8ff7 - Init COMPLETE +ip-26-0-172-252:1108193:1108692 [0] NCCL INFO comm 0xbab6300 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x721e6042a5cc8ff7 - Init COMPLETE +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO comm 0xbcaabe0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2f4ca0bf353b4f72 - Init START +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO comm 0xad3a390 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2f4ca0bf353b4f72 - Init START +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO comm 0xc38bbf0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x58336a8fe61562aa - Init START +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO comm 0xbacc560 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x58336a8fe61562aa - Init START +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/181 +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO comm 0xa7f6aa0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf1d4e1b0599776a7 - Init START +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO comm 0xbd52910 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf1d4e1b0599776a7 - Init START +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Using network Libfabric +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO comm 0xb5b11c0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3afcc559b307a251 - Init START +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO comm 0xc17c510 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3afcc559b307a251 - Init START +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/168 +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Connected all rings +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO Connected all trees +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Connected all rings +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO Connected all trees +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108199:1108723 [6] NCCL INFO comm 0xbcaabe0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2f4ca0bf353b4f72 - Init COMPLETE +ip-26-0-172-252:1108197:1108722 [4] NCCL INFO comm 0xad3a390 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2f4ca0bf353b4f72 - Init COMPLETE +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Connected all rings +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO Connected all trees +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Connected all rings +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO Connected all trees +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Connected all rings +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO Connected all trees +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Connected all rings +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO Connected all trees +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Connected all rings +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO Connected all trees +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Connected all rings +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO Connected all trees +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-252:1108200:1108732 [7] NCCL INFO comm 0xb5b11c0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3afcc559b307a251 - Init COMPLETE +ip-26-0-172-252:1108198:1108731 [5] NCCL INFO comm 0xc17c510 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3afcc559b307a251 - Init COMPLETE +ip-26-0-172-252:1108193:1108725 [0] NCCL INFO comm 0xbacc560 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x58336a8fe61562aa - Init COMPLETE +ip-26-0-172-252:1108195:1108726 [2] NCCL INFO comm 0xc38bbf0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x58336a8fe61562aa - Init COMPLETE +ip-26-0-172-252:1108194:1108728 [1] NCCL INFO comm 0xa7f6aa0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf1d4e1b0599776a7 - Init COMPLETE +ip-26-0-172-252:1108196:1108729 [3] NCCL INFO comm 0xbd52910 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf1d4e1b0599776a7 - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader)trainer.train(dataloader) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask)hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 3.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 1.38 GiB is free. Including non-PyTorch memory, this process has 77.94 GiB memory in use. Of the allocated memory 73.38 GiB is allocated by PyTorch, and 2.86 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + .contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 3.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 1.26 GiB is free. Including non-PyTorch memory, this process has 78.05 GiB memory in use. Of the allocated memory 73.38 GiB is allocated by PyTorch, and 2.86 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + .contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 3.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 1.50 GiB is free. Including non-PyTorch memory, this process has 77.82 GiB memory in use. Of the allocated memory 73.38 GiB is allocated by PyTorch, and 2.86 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + .contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 3.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 1.38 GiB is free. Including non-PyTorch memory, this process has 77.94 GiB memory in use. Of the allocated memory 73.38 GiB is allocated by PyTorch, and 2.86 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-172-252]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-172-252]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-172-252]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-172-252]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-172-252]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-172-252]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-172-252]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-172-252]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-172-252]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-172-252]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-172-252]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-172-252]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-172-252]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-172-252]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-172-252]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-172-252]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-172-252]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-172-252]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-172-252]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-172-252]: Max retries reached, giving up on communication +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-172-252]: Max retries reached, giving up on communication +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-172-252]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:23:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-172-252]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:23:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-172-252]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:23:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-172-252]: Max retries reached, giving up on communication +12/27/2024 17:23:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-172-252]: Max retries reached, giving up on communication +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + trainer.train(dataloader)trainer.train(dataloader) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque +[2024-12-27 17:24:07,229] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1108193) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 17:24:07,284] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic__ilzhou_/13803153_i4ovonti/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:23:41 + host : ip-26-0-172-252.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 1108194) + error_file: /tmp/torchelastic__ilzhou_/13803153_i4ovonti/attempt_0/1/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 3.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 1.38 GiB is free. Including non-PyTorch memory, this process has 77.94 GiB memory in use. Of the allocated memory 73.38 GiB is allocated by PyTorch, and 2.86 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[2]: + time : 2024-12-27_17:23:41 + host : ip-26-0-172-252.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 1108195) + error_file: /tmp/torchelastic__ilzhou_/13803153_i4ovonti/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 3.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 1.38 GiB is free. Including non-PyTorch memory, this process has 77.94 GiB memory in use. Of the allocated memory 73.38 GiB is allocated by PyTorch, and 2.86 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[3]: + time : 2024-12-27_17:23:41 + host : ip-26-0-172-252.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 1108196) + error_file: /tmp/torchelastic__ilzhou_/13803153_i4ovonti/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 3.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 1.26 GiB is free. Including non-PyTorch memory, this process has 78.05 GiB memory in use. Of the allocated memory 73.38 GiB is allocated by PyTorch, and 2.86 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[4]: + time : 2024-12-27_17:23:42 + host : ip-26-0-172-252.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 1108197) + error_file: /tmp/torchelastic__ilzhou_/13803153_i4ovonti/attempt_0/4/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +[5]: + time : 2024-12-27_17:23:42 + host : ip-26-0-172-252.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 1108198) + error_file: /tmp/torchelastic__ilzhou_/13803153_i4ovonti/attempt_0/5/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +[6]: + time : 2024-12-27_17:23:42 + host : ip-26-0-172-252.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 1108199) + error_file: /tmp/torchelastic__ilzhou_/13803153_i4ovonti/attempt_0/6/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +[7]: + time : 2024-12-27_17:23:42 + host : ip-26-0-172-252.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 1108200) + error_file: /tmp/torchelastic__ilzhou_/13803153_i4ovonti/attempt_0/7/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:23:41 + host : ip-26-0-172-252.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 1108193) + error_file: /tmp/torchelastic__ilzhou_/13803153_i4ovonti/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 3.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 1.50 GiB is free. Including non-PyTorch memory, this process has 77.82 GiB memory in use. Of the allocated memory 73.38 GiB is allocated by PyTorch, and 2.86 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +============================================================ +srun: error: ip-26-0-172-252: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13803153.0 diff --git a/logs/13803238-bench_3.57G_dp2_tp2_pp4_acc64_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13803238-bench_3.57G_dp2_tp2_pp4_acc64_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..b062671b9160700f64b541ca44d97179d0c49e22 --- /dev/null +++ b/logs/13803238-bench_3.57G_dp2_tp2_pp4_acc64_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,5103 @@ ++ '[' -z 13803238 ']' ++ unset FI_PROVIDER ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-161-[78,221]' ++ export 'NODELIST=ip-26-0-161-78 +ip-26-0-161-221' ++ NODELIST='ip-26-0-161-78 +ip-26-0-161-221' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-161-[78,221]' ++ export MASTER_NODE=ip-26-0-161-78 ++ MASTER_NODE=ip-26-0-161-78 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ export NCCL_P2P_LEVEL=LOC ++ NCCL_P2P_LEVEL=LOC ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-161-78' +Master node: ip-26-0-161-78 ++ echo 'All nodes: ip-26-0-161-78 +ip-26-0-161-221' +All nodes: ip-26-0-161-78 +ip-26-0-161-221 ++ echo 'World size: 16' +World size: 16 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13803238 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-161-78:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_3.57G_dp2_tp2_pp4_acc64_mbs2_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-27 17:25:06,907] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:25:06,914] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:25:06,907] torch.distributed.run: [WARNING] +[2024-12-27 17:25:06,907] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:25:06,907] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:25:06,907] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:25:06,914] torch.distributed.run: [WARNING] +[2024-12-27 17:25:06,914] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:25:06,914] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:25:06,914] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-161-221:2752205:2752205 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.161.221<0> +ip-26-0-161-221:2752205:2752205 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:2752205:2752205 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:2752205:2752205 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-78:1287866:1287866 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-78:1287864:1287864 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:2752206:2752206 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:2752208:2752208 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:2752212:2752212 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-78:1287866:1287866 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.161.78<0> +ip-26-0-161-78:1287864:1287864 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.161.78<0> +ip-26-0-161-78:1287866:1287866 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-78:1287866:1287866 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-78:1287864:1287864 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-78:1287864:1287864 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:2752206:2752206 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.161.221<0> +ip-26-0-161-221:2752212:2752212 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.161.221<0> +ip-26-0-161-221:2752208:2752208 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.161.221<0> +ip-26-0-161-221:2752206:2752206 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:2752206:2752206 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:2752212:2752212 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:2752212:2752212 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:2752209:2752209 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:2752211:2752211 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:2752208:2752208 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:2752208:2752208 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:2752209:2752209 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.161.221<0> +ip-26-0-161-221:2752211:2752211 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.161.221<0> +ip-26-0-161-78:1287868:1287868 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:2752210:2752210 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:2752211:2752211 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:2752209:2752209 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:2752211:2752211 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:2752209:2752209 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-78:1287865:1287865 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:2752210:2752210 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.161.221<0> +ip-26-0-161-78:1287868:1287868 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.161.78<0> +ip-26-0-161-221:2752210:2752210 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:2752210:2752210 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-78:1287865:1287865 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.161.78<0> +ip-26-0-161-78:1287868:1287868 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-78:1287868:1287868 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-78:1287865:1287865 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-78:1287865:1287865 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-78:1287861:1287861 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-78:1287861:1287861 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.161.78<0> +ip-26-0-161-78:1287863:1287863 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-78:1287861:1287861 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-78:1287861:1287861 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-78:1287863:1287863 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.161.78<0> +ip-26-0-161-78:1287863:1287863 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-78:1287863:1287863 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:2752207:2752207 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-78:1287862:1287862 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:2752207:2752207 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.161.221<0> +ip-26-0-161-78:1287862:1287862 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.161.78<0> +ip-26-0-161-221:2752207:2752207 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:2752207:2752207 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-78:1287862:1287862 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-78:1287862:1287862 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-78:1287867:1287867 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-78:1287867:1287867 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.161.78<0> +ip-26-0-161-78:1287867:1287867 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-78:1287867:1287867 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/205 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO comm 0x9294890 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO comm 0x8ac7860 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO comm 0x9fbbce0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO comm 0x9d658a0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO comm 0x857d660 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO comm 0x9e65d90 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO comm 0x911ab90 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO comm 0x88529b0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO comm 0x9a23d50 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO comm 0xa0b3f90 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO comm 0x8582c70 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO comm 0x8d3d080 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO comm 0xa0efb70 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO comm 0xa0404b0 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO comm 0x89d9230 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO comm 0x9a6ca30 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x254ac03d4fc985fe - Init START +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 7/-1/-1->15->-1 [3] 7/-1/-1->15->-1 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Trees [0] 4/12/-1->8->0 [1] 4/12/-1->8->0 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Trees [0] 8/-1/-1->0->-1 [1] 8/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Connected all rings +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Connected all rings +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Connected all rings +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Connected all rings +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Connected all rings +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Connected all rings +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Connected all rings +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Connected all rings +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Connected all rings +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Connected all rings +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Connected all rings +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Connected all rings +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Connected all rings +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Connected all rings +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Connected all rings +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Connected all rings +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO Connected all trees +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752205:2752367 [0] NCCL INFO comm 0xa0efb70 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO Connected all trees +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO Connected all trees +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO Connected all trees +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287867:1288029 [6] NCCL INFO comm 0x9294890 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO Connected all trees +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288025 [2] NCCL INFO comm 0x9e65d90 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-78:1287868:1288024 [7] NCCL INFO comm 0x8ac7860 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO Connected all trees +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287866:1288023 [5] NCCL INFO comm 0x9d658a0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-78:1287862:1288028 [1] NCCL INFO comm 0x88529b0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO Connected all trees +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO Connected all trees +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO Connected all trees +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287864:1288022 [3] NCCL INFO comm 0x857d660 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-78:1287865:1288026 [4] NCCL INFO comm 0x9fbbce0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO Connected all trees +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287861:1288027 [0] NCCL INFO comm 0x911ab90 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO Connected all trees +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752211:2752371 [6] NCCL INFO comm 0x8582c70 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO Connected all trees +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO Connected all trees +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752207:2752372 [2] NCCL INFO comm 0x89d9230 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-221:2752206:2752369 [1] NCCL INFO comm 0x8d3d080 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-221:2752210:2752374 [5] NCCL INFO comm 0xa0b3f90 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO Connected all trees +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO Connected all trees +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2752370 [7] NCCL INFO comm 0x9a6ca30 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-221:2752208:2752368 [3] NCCL INFO comm 0x9a23d50 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO Connected all trees +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752209:2752373 [4] NCCL INFO comm 0xa0404b0 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x254ac03d4fc985fe - Init COMPLETE +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO comm 0x8c10210 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO comm 0xa329590 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO comm 0x8f74760 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO comm 0x8cfe400 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO comm 0x94cad00 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO comm 0x9f9beb0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO comm 0xa1f2e70 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO comm 0xa09cb80 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO comm 0x87b4a80 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO comm 0x8a89180 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO comm 0x87b9400 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO comm 0x9ca4110 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO comm 0x93514e0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO comm 0xa2eac70 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO comm 0x9c5b3c0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO comm 0xa277580 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcbd75c61f55845d5 - Init START +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 7/-1/-1->15->-1 [3] 7/-1/-1->15->-1 +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Trees [0] 8/-1/-1->0->-1 [1] 8/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Trees [0] 4/12/-1->8->0 [1] 4/12/-1->8->0 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Connected all rings +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Connected all rings +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Connected all rings +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Connected all rings +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Connected all rings +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Connected all rings +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Connected all rings +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Connected all rings +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Connected all rings +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Connected all rings +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Connected all rings +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Connected all rings +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Connected all rings +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Connected all rings +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Connected all rings +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Connected all rings +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO Connected all trees +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287868:1288256 [7] NCCL INFO comm 0x8cfe400 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO Connected all trees +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO Connected all trees +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO Connected all trees +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752205:2752598 [0] NCCL INFO comm 0xa329590 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO Connected all trees +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288252 [2] NCCL INFO comm 0xa09cb80 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-78:1287867:1288257 [6] NCCL INFO comm 0x94cad00 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-78:1287866:1288254 [5] NCCL INFO comm 0x9f9beb0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO Connected all trees +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO Connected all trees +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752211:2752601 [6] NCCL INFO comm 0x87b9400 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO Connected all trees +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO Connected all trees +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752207:2752599 [2] NCCL INFO comm 0x8c10210 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-221:2752206:2752604 [1] NCCL INFO comm 0x8f74760 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO Connected all trees +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287864:1288253 [3] NCCL INFO comm 0x87b4a80 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO Connected all trees +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO Connected all trees +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287865:1288255 [4] NCCL INFO comm 0xa1f2e70 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-78:1287862:1288259 [1] NCCL INFO comm 0x8a89180 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-78:1287861:1288258 [0] NCCL INFO comm 0x93514e0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO Connected all trees +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO Connected all trees +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO Connected all trees +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752210:2752600 [5] NCCL INFO comm 0xa2eac70 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-221:2752212:2752603 [7] NCCL INFO comm 0x9ca4110 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-221:2752208:2752602 [3] NCCL INFO comm 0x9c5b3c0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xcbd75c61f55845d5 - Init COMPLETE +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO Connected all trees +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752209:2752605 [4] NCCL INFO comm 0xa277580 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcbd75c61f55845d5 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO comm 0x87cddc0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x84eae10f09f04031 - Init START +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO comm 0xa2ff1c0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x84eae10f09f04031 - Init START +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO comm 0x8a9d520 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa2629b57b7b82a55 - Init START +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO comm 0x8d12110 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa2629b57b7b82a55 - Init START +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO comm 0x9366bc0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa2629b57b7b82a55 - Init START +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO comm 0x87c8820 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa2629b57b7b82a55 - Init START +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO comm 0xa207080 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa2629b57b7b82a55 - Init START +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO comm 0xa0b1060 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa2629b57b7b82a55 - Init START +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO comm 0x9fb0650 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa2629b57b7b82a55 - Init START +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO comm 0x94df080 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa2629b57b7b82a55 - Init START +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO comm 0x8c24860 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x84eae10f09f04031 - Init START +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO comm 0xa28b990 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x84eae10f09f04031 - Init START +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO comm 0x9c6f600 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x84eae10f09f04031 - Init START +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO comm 0x8f88c60 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x84eae10f09f04031 - Init START +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO comm 0x9cb8980 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x84eae10f09f04031 - Init START +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO comm 0xa33dcc0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x84eae10f09f04031 - Init START +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Connected all rings +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Connected all rings +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Connected all rings +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Connected all rings +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Connected all rings +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Connected all rings +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Connected all rings +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Connected all rings +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO Connected all trees +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287868:1288292 [7] NCCL INFO comm 0x8d12110 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa2629b57b7b82a55 - Init COMPLETE +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO Connected all trees +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287861:1288286 [0] NCCL INFO comm 0x9366bc0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa2629b57b7b82a55 - Init COMPLETE +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO Connected all trees +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO Connected all trees +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO Connected all trees +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288287 [2] NCCL INFO comm 0xa0b1060 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa2629b57b7b82a55 - Init COMPLETE +ip-26-0-161-78:1287862:1288288 [1] NCCL INFO comm 0x8a9d520 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa2629b57b7b82a55 - Init COMPLETE +ip-26-0-161-78:1287867:1288293 [6] NCCL INFO comm 0x94df080 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa2629b57b7b82a55 - Init COMPLETE +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO Connected all trees +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287864:1288289 [3] NCCL INFO comm 0x87c8820 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa2629b57b7b82a55 - Init COMPLETE +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO Connected all trees +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO Connected all trees +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287866:1288290 [5] NCCL INFO comm 0x9fb0650 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa2629b57b7b82a55 - Init COMPLETE +ip-26-0-161-78:1287865:1288291 [4] NCCL INFO comm 0xa207080 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa2629b57b7b82a55 - Init COMPLETE +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Connected all rings +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Connected all rings +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Connected all rings +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Connected all rings +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Connected all rings +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Connected all rings +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Connected all rings +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Connected all rings +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO Connected all trees +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2752636 [7] NCCL INFO comm 0x9cb8980 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x84eae10f09f04031 - Init COMPLETE +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO Connected all trees +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752205:2752631 [0] NCCL INFO comm 0xa33dcc0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x84eae10f09f04031 - Init COMPLETE +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO Connected all trees +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752206:2752635 [1] NCCL INFO comm 0x8f88c60 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x84eae10f09f04031 - Init COMPLETE +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO Connected all trees +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO Connected all trees +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752211:2752637 [6] NCCL INFO comm 0x87cddc0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x84eae10f09f04031 - Init COMPLETE +ip-26-0-161-221:2752207:2752634 [2] NCCL INFO comm 0x8c24860 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x84eae10f09f04031 - Init COMPLETE +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO Connected all trees +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752208:2752633 [3] NCCL INFO comm 0x9c6f600 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x84eae10f09f04031 - Init COMPLETE +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO Connected all trees +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO Connected all trees +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752210:2752632 [5] NCCL INFO comm 0xa2ff1c0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x84eae10f09f04031 - Init COMPLETE +ip-26-0-161-221:2752209:2752638 [4] NCCL INFO comm 0xa28b990 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x84eae10f09f04031 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO comm 0xbf079a0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x272136be6388bdd3 - Init START +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO comm 0xbf7b5a0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x272136be6388bdd3 - Init START +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO comm 0xa449b60 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x427efc08b0f5056b - Init START +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO comm 0xb935330 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x427efc08b0f5056b - Init START +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO comm 0xbd2b360 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x40bfbc7e3a4579ef - Init START +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO comm 0xafe1ea0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc254f12f49967e6c - Init START +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO comm 0xa4433a0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x40bfbc7e3a4579ef - Init START +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO comm 0xa719080 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc254f12f49967e6c - Init START +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO comm 0xaa64040 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xab42ffc0d73e403e - Init START +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO comm 0xbaad320 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xab42ffc0d73e403e - Init START +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO comm 0xbc25b70 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x60a6d26916eb8fab - Init START +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO comm 0xbe7c2f0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x60a6d26916eb8fab - Init START +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO comm 0xadc5cb0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8b48c07babd4a67c - Init START +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO comm 0xc17f180 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8b48c07babd4a67c - Init START +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO comm 0xa986420 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xaecec93958cc6b26 - Init START +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO comm 0xb154470 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xaecec93958cc6b26 - Init START +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Connected all rings +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO Connected all trees +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Connected all rings +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO Connected all trees +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287865:1288349 [4] NCCL INFO comm 0xbe7c2f0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x60a6d26916eb8fab - Init COMPLETE +ip-26-0-161-78:1287866:1288350 [5] NCCL INFO comm 0xbc25b70 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x60a6d26916eb8fab - Init COMPLETE +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Connected all rings +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO Connected all trees +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Connected all rings +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO Connected all trees +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287868:1288351 [7] NCCL INFO comm 0xa986420 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xaecec93958cc6b26 - Init COMPLETE +ip-26-0-161-78:1287867:1288348 [6] NCCL INFO comm 0xb154470 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xaecec93958cc6b26 - Init COMPLETE +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Connected all rings +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO Connected all trees +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Connected all rings +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO Connected all trees +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Connected all rings +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO Connected all trees +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Connected all rings +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO Connected all trees +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Connected all rings +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO Connected all trees +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287864:1288333 [3] NCCL INFO comm 0xa4433a0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x40bfbc7e3a4579ef - Init COMPLETE +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Connected all rings +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO Connected all trees +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287862:1288330 [1] NCCL INFO comm 0xa719080 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc254f12f49967e6c - Init COMPLETE +ip-26-0-161-78:1287863:1288332 [2] NCCL INFO comm 0xbd2b360 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x40bfbc7e3a4579ef - Init COMPLETE +ip-26-0-161-78:1287861:1288329 [0] NCCL INFO comm 0xafe1ea0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc254f12f49967e6c - Init COMPLETE +ip-26-0-161-221:2752209:2752673 [4] NCCL INFO comm 0xbf079a0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x272136be6388bdd3 - Init COMPLETE +ip-26-0-161-221:2752210:2752674 [5] NCCL INFO comm 0xbf7b5a0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x272136be6388bdd3 - Init COMPLETE +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Connected all rings +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO Connected all trees +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Connected all rings +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO Connected all trees +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Connected all rings +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO Connected all trees +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Connected all rings +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO Connected all trees +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Connected all rings +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO Connected all trees +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752206:2752695 [1] NCCL INFO comm 0xadc5cb0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8b48c07babd4a67c - Init COMPLETE +ip-26-0-161-221:2752208:2752692 [3] NCCL INFO comm 0xbaad320 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xab42ffc0d73e403e - Init COMPLETE +ip-26-0-161-221:2752207:2752691 [2] NCCL INFO comm 0xaa64040 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xab42ffc0d73e403e - Init COMPLETE +ip-26-0-161-221:2752205:2752694 [0] NCCL INFO comm 0xc17f180 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8b48c07babd4a67c - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Connected all rings +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO Connected all trees +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO comm 0xbc38830 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeaea8258c4777480 - Init START +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO comm 0xbf8e260 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeaea8258c4777480 - Init START +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO comm 0xa72bd40 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeaea8258c4777480 - Init START +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO comm 0xadd8e80 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeaea8258c4777480 - Init START +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO comm 0xbf1a660 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x3382b0d502f958c3 - Init START +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO comm 0xbe8efb0 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x3382b0d502f958c3 - Init START +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO comm 0xaff4b60 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3382b0d502f958c3 - Init START +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO comm 0xc191e40 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3382b0d502f958c3 - Init START +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752212:2752681 [7] NCCL INFO comm 0xb935330 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x427efc08b0f5056b - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752211:2752680 [6] NCCL INFO comm 0xa449b60 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x427efc08b0f5056b - Init COMPLETE +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO comm 0xa9990e0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf92c32d4261b7834 - Init START +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO comm 0xbabffe0 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf92c32d4261b7834 - Init START +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO comm 0xb947ff0 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf92c32d4261b7834 - Init START +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO comm 0xa456060 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf92c32d4261b7834 - Init START +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO comm 0xb167130 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8727a4bb6d332526 - Init START +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO comm 0xbd3e020 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8727a4bb6d332526 - Init START +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO comm 0xa45c820 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8727a4bb6d332526 - Init START +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO comm 0xaa76d00 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8727a4bb6d332526 - Init START +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 00/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 01/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 02/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 03/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 00/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 01/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 02/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 03/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 00/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 01/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 02/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 03/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 00/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 00/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 01/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 01/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 02/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 02/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 03/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 03/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 00/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 01/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 02/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 03/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 00/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 01/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 02/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 03/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 00/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 01/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 02/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 00/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 03/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 01/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 02/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 03/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 00/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 01/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 02/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 03/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 00/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 01/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 02/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 03/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 00/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 01/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 02/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 03/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 00/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 01/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 02/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 03/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Connected all rings +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Connected all rings +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Connected all rings +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 00/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 01/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 02/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 03/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Connected all rings +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 00/0 : 3[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 01/0 : 3[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Connected all rings +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Connected all rings +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Connected all rings +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Connected all rings +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Connected all rings +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Connected all rings +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Connected all rings +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 00/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 01/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 02/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 03/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Connected all rings +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 00/0 : 3[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 01/0 : 3[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Connected all rings +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 00/0 : 3[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 01/0 : 3[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Connected all rings +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 00/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 01/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 02/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 03/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Connected all rings +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 00/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 01/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 02/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 03/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Connected all rings +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 00/0 : 3[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 01/0 : 3[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO Connected all trees +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287868:1288374 [7] NCCL INFO comm 0xa9990e0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf92c32d4261b7834 - Init COMPLETE +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO Connected all trees +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752208:2752723 [3] NCCL INFO comm 0xbabffe0 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf92c32d4261b7834 - Init COMPLETE +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO Connected all trees +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO Connected all trees +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287864:1288375 [3] NCCL INFO comm 0xa456060 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf92c32d4261b7834 - Init COMPLETE +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO Connected all trees +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287865:1288372 [4] NCCL INFO comm 0xbe8efb0 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x3382b0d502f958c3 - Init COMPLETE +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO Connected all trees +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO Connected all trees +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO Connected all trees +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287866:1288370 [5] NCCL INFO comm 0xbc38830 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeaea8258c4777480 - Init COMPLETE +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO Connected all trees +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287867:1288377 [6] NCCL INFO comm 0xb167130 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8727a4bb6d332526 - Init COMPLETE +ip-26-0-161-221:2752205:2752719 [0] NCCL INFO comm 0xc191e40 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3382b0d502f958c3 - Init COMPLETE +ip-26-0-161-221:2752206:2752716 [1] NCCL INFO comm 0xadd8e80 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeaea8258c4777480 - Init COMPLETE +ip-26-0-161-221:2752207:2752726 [2] NCCL INFO comm 0xaa76d00 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8727a4bb6d332526 - Init COMPLETE +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO Connected all trees +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO Connected all trees +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO Connected all trees +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO Connected all trees +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2752724 [7] NCCL INFO comm 0xb947ff0 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf92c32d4261b7834 - Init COMPLETE +ip-26-0-161-78:1287862:1288371 [1] NCCL INFO comm 0xa72bd40 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeaea8258c4777480 - Init COMPLETE +ip-26-0-161-78:1287863:1288376 [2] NCCL INFO comm 0xbd3e020 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8727a4bb6d332526 - Init COMPLETE +ip-26-0-161-78:1287861:1288373 [0] NCCL INFO comm 0xaff4b60 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3382b0d502f958c3 - Init COMPLETE +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO Connected all trees +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO Connected all trees +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO Connected all trees +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752210:2752717 [5] NCCL INFO comm 0xbf8e260 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeaea8258c4777480 - Init COMPLETE +ip-26-0-161-221:2752209:2752720 [4] NCCL INFO comm 0xbf1a660 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x3382b0d502f958c3 - Init COMPLETE +ip-26-0-161-221:2752211:2752727 [6] NCCL INFO comm 0xa45c820 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8727a4bb6d332526 - Init COMPLETE +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO comm 0xbad6ae0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x10152fa7d2232417 - Init START +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO comm 0xadf0490 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x10152fa7d2232417 - Init START +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO comm 0xc1a65c0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9053c34125b578d7 - Init START +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO comm 0xaa8d290 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9053c34125b578d7 - Init START +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO comm 0xa9021a0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa0bb9d6feb45c0bf - Init START +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO comm 0xa62d4e0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa0bb9d6feb45c0bf - Init START +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO comm 0xab71fa0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3ed331340df6d18b - Init START +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO comm 0xbe10820 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3ed331340df6d18b - Init START +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO comm 0xc164410 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa589ecab132bb476 - Init START +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO comm 0xbb1d990 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa589ecab132bb476 - Init START +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO comm 0xb1caff0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4ca719230bac3919 - Init START +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO comm 0xbf16120 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4ca719230bac3919 - Init START +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO comm 0xc066f90 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4e68b3425f8ce6ed - Init START +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO comm 0xb340770 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4e68b3425f8ce6ed - Init START +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO comm 0xc0f0250 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc8f2cfd6fdd4bfd4 - Init START +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO comm 0xa633510 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc8f2cfd6fdd4bfd4 - Init START +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Connected all rings +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO Connected all trees +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Connected all rings +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO Connected all trees +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-221:2752206:2752748 [1] NCCL INFO comm 0xadf0490 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x10152fa7d2232417 - Init COMPLETE +ip-26-0-161-221:2752208:2752749 [3] NCCL INFO comm 0xbad6ae0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x10152fa7d2232417 - Init COMPLETE +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Connected all rings +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO Connected all trees +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Connected all rings +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO Connected all trees +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Connected all rings +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO Connected all trees +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Connected all rings +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO Connected all trees +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752205:2752751 [0] NCCL INFO comm 0xc1a65c0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9053c34125b578d7 - Init COMPLETE +ip-26-0-161-221:2752207:2752752 [2] NCCL INFO comm 0xaa8d290 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9053c34125b578d7 - Init COMPLETE +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Connected all rings +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO Connected all trees +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Connected all rings +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO Connected all trees +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2752755 [7] NCCL INFO comm 0xbb1d990 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa589ecab132bb476 - Init COMPLETE +ip-26-0-161-221:2752211:2752757 [6] NCCL INFO comm 0xa633510 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc8f2cfd6fdd4bfd4 - Init COMPLETE +ip-26-0-161-221:2752209:2752758 [4] NCCL INFO comm 0xc0f0250 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc8f2cfd6fdd4bfd4 - Init COMPLETE +ip-26-0-161-221:2752210:2752754 [5] NCCL INFO comm 0xc164410 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa589ecab132bb476 - Init COMPLETE +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Connected all rings +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO Connected all trees +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Connected all rings +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO Connected all trees +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287864:1288399 [3] NCCL INFO comm 0xa62d4e0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa0bb9d6feb45c0bf - Init COMPLETE +ip-26-0-161-78:1287862:1288398 [1] NCCL INFO comm 0xa9021a0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa0bb9d6feb45c0bf - Init COMPLETE +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Connected all rings +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO Connected all trees +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Connected all rings +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO Connected all trees +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Connected all rings +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO Connected all trees +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287867:1288407 [6] NCCL INFO comm 0xb340770 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4e68b3425f8ce6ed - Init COMPLETE +ip-26-0-161-78:1287865:1288405 [4] NCCL INFO comm 0xc066f90 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4e68b3425f8ce6ed - Init COMPLETE +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Connected all rings +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO Connected all trees +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Connected all rings +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO Connected all trees +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287868:1288402 [7] NCCL INFO comm 0xab71fa0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3ed331340df6d18b - Init COMPLETE +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Connected all rings +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO Connected all trees +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288408 [2] NCCL INFO comm 0xbf16120 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4ca719230bac3919 - Init COMPLETE +ip-26-0-161-78:1287861:1288406 [0] NCCL INFO comm 0xb1caff0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4ca719230bac3919 - Init COMPLETE +ip-26-0-161-78:1287866:1288401 [5] NCCL INFO comm 0xbe10820 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3ed331340df6d18b - Init COMPLETE +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO comm 0xbe389e0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd4ab86e74624544f - Init START +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO comm 0xc08e680 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x77e4ef292d37c0b5 - Init START +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO comm 0xae2a2d0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd4ab86e74624544f - Init START +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO comm 0xc1ded40 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x77e4ef292d37c0b5 - Init START +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO comm 0xb368cc0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4f6ac561ce7c43f6 - Init START +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO comm 0xaac5670 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4f6ac561ce7c43f6 - Init START +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO comm 0xab993c0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa4a10987024f590b - Init START +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO comm 0xbb0f600 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa4a10987024f590b - Init START +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Connected all rings +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO Connected all trees +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287865:1288428 [4] NCCL INFO comm 0xc08e680 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x77e4ef292d37c0b5 - Init COMPLETE +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Connected all rings +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO Connected all trees +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Connected all rings +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO Connected all trees +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Connected all rings +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO Connected all trees +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Connected all rings +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO Connected all trees +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752205:2752779 [0] NCCL INFO comm 0xc1ded40 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x77e4ef292d37c0b5 - Init COMPLETE +ip-26-0-161-221:2752208:2752784 [3] NCCL INFO comm 0xbb0f600 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa4a10987024f590b - Init COMPLETE +ip-26-0-161-221:2752207:2752782 [2] NCCL INFO comm 0xaac5670 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4f6ac561ce7c43f6 - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-221:2752206:2752780 [1] NCCL INFO comm 0xae2a2d0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd4ab86e74624544f - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Connected all rings +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO Connected all trees +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Connected all rings +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO Connected all trees +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Connected all rings +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO Connected all trees +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287867:1288430 [6] NCCL INFO comm 0xb368cc0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4f6ac561ce7c43f6 - Init COMPLETE +ip-26-0-161-78:1287868:1288431 [7] NCCL INFO comm 0xab993c0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa4a10987024f590b - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-78:1287866:1288429 [5] NCCL INFO comm 0xbe389e0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd4ab86e74624544f - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO comm 0xc44cb10 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5f12d1e3b1ac02d7 - Init START +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO comm 0x1b5f2a50 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5f12d1e3b1ac02d7 - Init START +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO comm 0xc35aa20 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x1b623cbdb17b08c1 - Init START +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO comm 0x1cd3a170 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1b623cbdb17b08c1 - Init START +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO comm 0xbdd5500 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9d1242bd56c46bf6 - Init START +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO comm 0x1ce13a40 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9d1242bd56c46bf6 - Init START +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO comm 0x1b419f00 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xbe77f6eb5be0e052 - Init START +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO comm 0xa9923f0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbe77f6eb5be0e052 - Init START +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Connected all rings +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO Connected all trees +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Connected all rings +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO Connected all trees +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Connected all rings +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO Connected all trees +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Connected all rings +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO Connected all trees +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752206:2752806 [1] NCCL INFO comm 0x1b5f2a50 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5f12d1e3b1ac02d7 - Init COMPLETE +ip-26-0-161-221:2752206:2752835 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-221:2752206:2752835 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-221:2752210:2752807 [5] NCCL INFO comm 0xc44cb10 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5f12d1e3b1ac02d7 - Init COMPLETE +ip-26-0-161-221:2752210:2752836 [5] NCCL INFO Channel 02/1 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-221:2752210:2752836 [5] NCCL INFO Channel 03/1 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-221:2752205:2752808 [0] NCCL INFO comm 0x1cd3a170 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1b623cbdb17b08c1 - Init COMPLETE +ip-26-0-161-221:2752209:2752809 [4] NCCL INFO comm 0xc35aa20 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x1b623cbdb17b08c1 - Init COMPLETE +ip-26-0-161-221:2752205:2752837 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-221:2752205:2752837 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-221:2752209:2752838 [4] NCCL INFO Channel 02/1 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-221:2752209:2752838 [4] NCCL INFO Channel 03/1 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Connected all rings +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO Connected all trees +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Connected all rings +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO Connected all trees +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752208:2752820 [3] NCCL INFO comm 0x1ce13a40 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9d1242bd56c46bf6 - Init COMPLETE +ip-26-0-161-221:2752212:2752823 [7] NCCL INFO comm 0xbdd5500 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9d1242bd56c46bf6 - Init COMPLETE +ip-26-0-161-221:2752208:2752839 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-221:2752208:2752839 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-221:2752212:2752840 [7] NCCL INFO Channel 02/1 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-221:2752212:2752840 [7] NCCL INFO Channel 03/1 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Connected all rings +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO Connected all trees +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Connected all rings +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO Connected all trees +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752207:2752822 [2] NCCL INFO comm 0x1b419f00 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xbe77f6eb5be0e052 - Init COMPLETE +ip-26-0-161-221:2752207:2752841 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-221:2752211:2752824 [6] NCCL INFO comm 0xa9923f0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbe77f6eb5be0e052 - Init COMPLETE +ip-26-0-161-221:2752207:2752841 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-221:2752211:2752842 [6] NCCL INFO Channel 02/1 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-221:2752211:2752842 [6] NCCL INFO Channel 03/1 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO comm 0xadece80 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x10db0d36b148d148 - Init START +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO comm 0x16af0c50 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x10db0d36b148d148 - Init START +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO comm 0xb6b6550 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcf9629292637a53d - Init START +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO comm 0x14e85410 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcf9629292637a53d - Init START +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Channel 00/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Channel 01/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Channel 02/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Channel 03/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 00/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 01/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 02/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 03/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Channel 00/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Channel 01/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Channel 02/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Channel 03/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 00/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 01/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 02/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 03/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO comm 0x1462ec90 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x176e50e800b062b6 - Init START +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO comm 0xa8dcab0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x176e50e800b062b6 - Init START +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO comm 0xc1ec6b0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3596391d102b9ac5 - Init START +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO comm 0x14a44a90 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x3596391d102b9ac5 - Init START +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Connected all rings +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO Connected all trees +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Channel 00/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Channel 01/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Channel 02/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Channel 03/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 00/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Connected all rings +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO Connected all trees +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 01/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 02/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 03/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Channel 00/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Channel 01/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Channel 02/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Channel 03/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Connected all rings +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO Connected all trees +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 00/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 01/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 02/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 03/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287862:1288458 [1] NCCL INFO comm 0xadece80 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x10db0d36b148d148 - Init COMPLETE +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Connected all rings +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO Connected all trees +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287862:1288470 [1] NCCL INFO Channel 02/1 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-78:1287862:1288470 [1] NCCL INFO Channel 03/1 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-221:2752210:2752855 [5] NCCL INFO comm 0x16af0c50 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x10db0d36b148d148 - Init COMPLETE +ip-26-0-161-221:2752210:2752870 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-221:2752210:2752870 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-78:1287861:1288459 [0] NCCL INFO comm 0xb6b6550 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcf9629292637a53d - Init COMPLETE +ip-26-0-161-78:1287861:1288471 [0] NCCL INFO Channel 02/1 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-78:1287861:1288471 [0] NCCL INFO Channel 03/1 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-221:2752209:2752856 [4] NCCL INFO comm 0x14e85410 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcf9629292637a53d - Init COMPLETE +ip-26-0-161-221:2752209:2752871 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-221:2752209:2752871 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Connected all rings +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO Connected all trees +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288465 [2] NCCL INFO comm 0xc1ec6b0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3596391d102b9ac5 - Init COMPLETE +ip-26-0-161-78:1287863:1288474 [2] NCCL INFO Channel 02/1 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-78:1287863:1288474 [2] NCCL INFO Channel 03/1 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Connected all rings +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO Connected all trees +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Connected all rings +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO Connected all trees +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Connected all rings +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO Connected all trees +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2752864 [7] NCCL INFO comm 0x1462ec90 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x176e50e800b062b6 - Init COMPLETE +ip-26-0-161-221:2752212:2752874 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-221:2752212:2752874 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-78:1287864:1288464 [3] NCCL INFO comm 0xa8dcab0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x176e50e800b062b6 - Init COMPLETE +ip-26-0-161-78:1287864:1288475 [3] NCCL INFO Channel 02/1 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-78:1287864:1288475 [3] NCCL INFO Channel 03/1 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-221:2752211:2752863 [6] NCCL INFO comm 0x14a44a90 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x3596391d102b9ac5 - Init COMPLETE +ip-26-0-161-221:2752211:2752875 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-221:2752211:2752875 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO comm 0xcadc110 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x850b662679a9c0db - Init START +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO comm 0x13d302c0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x850b662679a9c0db - Init START +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO comm 0xc866d00 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2cf73c59a2f84d76 - Init START +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO comm 0x13571560 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2cf73c59a2f84d76 - Init START +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO comm 0x14bb5e60 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2aac5d9d692653c5 - Init START +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO comm 0xbd99530 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2aac5d9d692653c5 - Init START +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO comm 0xb5c8130 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x51eb34821e02671e - Init START +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO comm 0x12e0dcc0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x51eb34821e02671e - Init START +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Connected all rings +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO Connected all trees +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287866:1288492 [5] NCCL INFO comm 0xc866d00 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2cf73c59a2f84d76 - Init COMPLETE +ip-26-0-161-78:1287866:1288516 [5] NCCL INFO Channel 02/1 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-78:1287866:1288516 [5] NCCL INFO Channel 03/1 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Connected all rings +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO Connected all trees +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Connected all rings +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO Connected all trees +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Connected all rings +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO Connected all trees +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287865:1288491 [4] NCCL INFO comm 0xcadc110 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x850b662679a9c0db - Init COMPLETE +ip-26-0-161-78:1287862:1288490 [1] NCCL INFO comm 0x13571560 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2cf73c59a2f84d76 - Init COMPLETE +ip-26-0-161-78:1287865:1288518 [4] NCCL INFO Channel 02/1 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-78:1287862:1288517 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-78:1287862:1288517 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-78:1287865:1288518 [4] NCCL INFO Channel 03/1 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-78:1287861:1288489 [0] NCCL INFO comm 0x13d302c0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x850b662679a9c0db - Init COMPLETE +ip-26-0-161-78:1287861:1288519 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-78:1287861:1288519 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Connected all rings +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO Connected all trees +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Connected all rings +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO Connected all trees +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287867:1288505 [6] NCCL INFO comm 0xbd99530 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2aac5d9d692653c5 - Init COMPLETE +ip-26-0-161-78:1287867:1288521 [6] NCCL INFO Channel 02/1 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-78:1287867:1288521 [6] NCCL INFO Channel 03/1 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-78:1287863:1288503 [2] NCCL INFO comm 0x14bb5e60 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2aac5d9d692653c5 - Init COMPLETE +ip-26-0-161-78:1287863:1288522 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-78:1287863:1288522 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Connected all rings +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO Connected all trees +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Connected all rings +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO Connected all trees +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287868:1288507 [7] NCCL INFO comm 0xb5c8130 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x51eb34821e02671e - Init COMPLETE +ip-26-0-161-78:1287868:1288523 [7] NCCL INFO Channel 02/1 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-78:1287868:1288523 [7] NCCL INFO Channel 03/1 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-78:1287864:1288506 [3] NCCL INFO comm 0x12e0dcc0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x51eb34821e02671e - Init COMPLETE +ip-26-0-161-78:1287864:1288524 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-78:1287864:1288524 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-78:1287861:1288601 [0] NCCL INFO Channel 02/1 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-78:1287861:1288601 [0] NCCL INFO Channel 03/1 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-78:1287862:1288600 [1] NCCL INFO Channel 02/1 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-78:1287862:1288600 [1] NCCL INFO Channel 03/1 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-78:1287865:1288602 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-78:1287865:1288602 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-78:1287866:1288603 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-78:1287866:1288603 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-78:1287863:1288605 [2] NCCL INFO Channel 02/1 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-78:1287863:1288605 [2] NCCL INFO Channel 03/1 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-78:1287864:1288606 [3] NCCL INFO Channel 02/1 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-78:1287864:1288606 [3] NCCL INFO Channel 03/1 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-78:1287868:1288607 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-78:1287868:1288607 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-78:1287867:1288608 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-78:1287867:1288608 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-221:2752210:2752960 [5] NCCL INFO Channel 02/1 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-221:2752209:2752961 [4] NCCL INFO Channel 02/1 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-221:2752209:2752961 [4] NCCL INFO Channel 03/1 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-221:2752210:2752960 [5] NCCL INFO Channel 03/1 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-78:1287862:1288610 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-78:1287861:1288611 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-78:1287861:1288611 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-78:1287862:1288610 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-221:2752211:2752963 [6] NCCL INFO Channel 02/1 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-221:2752211:2752963 [6] NCCL INFO Channel 03/1 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-221:2752212:2752964 [7] NCCL INFO Channel 02/1 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-221:2752212:2752964 [7] NCCL INFO Channel 03/1 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-78:1287863:1288613 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-78:1287863:1288613 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-78:1287864:1288614 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-78:1287864:1288614 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-221:2752205:2752966 [0] NCCL INFO Channel 02/1 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-221:2752205:2752966 [0] NCCL INFO Channel 03/1 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-221:2752206:2752967 [1] NCCL INFO Channel 02/1 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-221:2752206:2752967 [1] NCCL INFO Channel 03/1 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-221:2752210:2752968 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-221:2752210:2752968 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-221:2752209:2752969 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-221:2752209:2752969 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-221:2752208:2752971 [3] NCCL INFO Channel 02/1 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-221:2752207:2752970 [2] NCCL INFO Channel 02/1 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-221:2752208:2752971 [3] NCCL INFO Channel 03/1 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-221:2752207:2752970 [2] NCCL INFO Channel 03/1 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-221:2752211:2752973 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-221:2752211:2752973 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-221:2752212:2752974 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-221:2752212:2752974 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA/Shared +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Using network Libfabric +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO comm 0x20894d10 rank 7 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x22aade7ce12590c4 - Init START +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO comm 0x2044f5f0 rank 6 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x22aade7ce12590c4 - Init START +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO comm 0x14b86cc0 rank 5 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x22aade7ce12590c4 - Init START +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO comm 0x181847c0 rank 3 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x22aade7ce12590c4 - Init START +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO comm 0x1e45d690 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x22aade7ce12590c4 - Init START +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO comm 0x16517cf0 rank 2 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x22aade7ce12590c4 - Init START +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO comm 0x15347dc0 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x22aade7ce12590c4 - Init START +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO comm 0x1cd1c7a0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x22aade7ce12590c4 - Init START +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO comm 0x1fb84f10 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd54467debc3446a1 - Init START +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO comm 0x1f9acc50 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd54467debc3446a1 - Init START +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO comm 0x1cb66dc0 rank 0 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd54467debc3446a1 - Init START +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO comm 0x14428510 rank 5 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd54467debc3446a1 - Init START +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO comm 0x15cc1e20 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd54467debc3446a1 - Init START +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO comm 0x1e55d850 rank 1 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd54467debc3446a1 - Init START +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO comm 0x160d6b60 rank 2 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd54467debc3446a1 - Init START +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO comm 0x161d2c40 rank 4 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd54467debc3446a1 - Init START +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 00/0 : 4[0] -> 5[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 01/0 : 4[0] -> 5[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 02/0 : 4[0] -> 5[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 03/0 : 4[0] -> 5[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 00/0 : 5[1] -> 6[4] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 01/0 : 5[1] -> 6[4] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 02/0 : 5[1] -> 6[4] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 03/0 : 5[1] -> 6[4] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 00/0 : 3[5] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 01/0 : 3[5] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 00/0 : 5[1] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 02/0 : 3[5] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 03/0 : 3[5] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 01/0 : 5[1] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 00/0 : 4[0] -> 5[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 02/0 : 5[1] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 01/0 : 4[0] -> 5[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 03/0 : 5[1] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 02/0 : 4[0] -> 5[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 00/0 : 6[4] -> 7[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 03/0 : 4[0] -> 5[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 01/0 : 6[4] -> 7[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 00/0 : 7[5] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 02/0 : 6[4] -> 7[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 01/0 : 7[5] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 03/0 : 6[4] -> 7[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 02/0 : 7[5] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 03/0 : 7[5] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 00/0 : 6[4] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 01/0 : 6[4] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 02/0 : 6[4] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 03/0 : 6[4] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 00/0 : 7[5] -> 0[0] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 01/0 : 7[5] -> 0[0] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 02/0 : 7[5] -> 0[0] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 03/0 : 7[5] -> 0[0] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 00/0 : 1[1] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 01/0 : 1[1] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 02/0 : 1[1] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 03/0 : 1[1] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 00/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 01/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 02/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 03/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 00/0 : 3[5] -> 4[0] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 01/0 : 3[5] -> 4[0] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 02/0 : 3[5] -> 4[0] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 03/0 : 3[5] -> 4[0] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[4] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[4] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[4] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[4] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 00/0 : 5[3] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 01/0 : 5[3] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 02/0 : 5[3] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 03/0 : 5[3] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 00/0 : 3[7] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 00/0 : 4[2] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 01/0 : 3[7] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 02/0 : 3[7] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 01/0 : 4[2] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 03/0 : 3[7] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 02/0 : 4[2] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 00/0 : 4[2] -> 5[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 03/0 : 4[2] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 01/0 : 4[2] -> 5[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 00/0 : 5[3] -> 6[6] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 02/0 : 4[2] -> 5[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 01/0 : 5[3] -> 6[6] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 03/0 : 4[2] -> 5[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 02/0 : 5[3] -> 6[6] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 03/0 : 5[3] -> 6[6] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[2] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[2] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[2] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[2] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 00/0 : 7[7] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 01/0 : 7[7] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 00/0 : 1[3] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 02/0 : 7[7] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 03/0 : 7[7] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 01/0 : 1[3] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 02/0 : 1[3] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 03/0 : 1[3] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[6] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[6] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[6] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[6] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 00/0 : 3[7] -> 4[2] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 01/0 : 3[7] -> 4[2] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 02/0 : 3[7] -> 4[2] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 03/0 : 3[7] -> 4[2] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Connected all rings +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 02/0 : 3[5] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 03/0 : 3[5] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 02/0 : 7[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 03/0 : 7[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Connected all rings +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 00/0 : 4[0] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 01/0 : 4[0] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Connected all rings +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 02/0 : 3[5] -> 5[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 03/0 : 3[5] -> 5[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Connected all rings +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 00/0 : 2[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 01/0 : 2[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 00/0 : 4[0] -> 6[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 01/0 : 4[0] -> 6[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 00/0 : 6[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 01/0 : 6[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Connected all rings +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 02/0 : 3[7] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 03/0 : 3[7] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Connected all rings +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 00/0 : 4[2] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 01/0 : 4[2] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Connected all rings +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Connected all rings +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 00/0 : 2[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 01/0 : 2[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 00/0 : 4[2] -> 6[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 01/0 : 4[2] -> 6[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Connected all rings +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Connected all rings +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Connected all rings +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 00/0 : 2[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 01/0 : 2[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 00/0 : 4[0] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 01/0 : 4[0] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Connected all rings +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 02/0 : 1[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 03/0 : 1[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 02/0 : 3[5] -> 5[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 03/0 : 3[5] -> 5[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 00/0 : 6[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 02/0 : 3[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 01/0 : 6[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 00/0 : 4[0] -> 2[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 03/0 : 3[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 02/0 : 5[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 01/0 : 4[0] -> 2[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 03/0 : 5[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 02/0 : 7[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 03/0 : 7[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 02/0 : 3[5] -> 7[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 03/0 : 3[5] -> 7[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 00/0 : 7[5] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 02/0 : 5[1] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 00/0 : 3[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Channel 03/0 : 5[1] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 01/0 : 3[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 01/0 : 7[5] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 02/0 : 5[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[1] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 00/0 : 7[5] -> 6[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 00/0 : 6[4] -> 5[1] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 03/0 : 5[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 01/0 : 6[4] -> 5[1] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Channel 01/0 : 7[5] -> 6[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[1] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 02/0 : 6[4] -> 5[1] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[1] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[1] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Channel 03/0 : 6[4] -> 5[1] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 00/0 : 2[4] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 01/0 : 2[4] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 02/0 : 2[4] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 03/0 : 2[4] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 00/0 : 6[4] -> 5[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 01/0 : 6[4] -> 5[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 02/0 : 6[4] -> 5[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 03/0 : 6[4] -> 5[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 02/0 : 5[1] -> 4[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Channel 03/0 : 5[1] -> 4[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Connected all rings +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 00/0 : 2[6] -> 4[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 01/0 : 2[6] -> 4[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 00/0 : 4[2] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 00/0 : 0[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 01/0 : 4[2] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 01/0 : 0[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Connected all rings +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Connected all rings +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 00/0 : 4[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 01/0 : 4[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 00/0 : 4[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 00/0 : 0[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 01/0 : 4[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 01/0 : 0[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 00/0 : 6[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 01/0 : 6[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 00/0 : 4[2] -> 2[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 01/0 : 4[2] -> 2[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 02/0 : 5[3] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Channel 03/0 : 5[3] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[3] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[3] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[3] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[3] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[3] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[3] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[3] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[3] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Connected all rings +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 02/0 : 1[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 03/0 : 1[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 02/0 : 3[7] -> 5[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 03/0 : 3[7] -> 5[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 02/0 : 3[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 03/0 : 3[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 02/0 : 5[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 03/0 : 5[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 02/0 : 5[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 03/0 : 5[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 00/0 : 6[6] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 01/0 : 6[6] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 00/0 : 2[6] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 02/0 : 6[6] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 03/0 : 6[6] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 01/0 : 2[6] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 02/0 : 2[6] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 02/0 : 5[3] -> 4[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 03/0 : 2[6] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Channel 03/0 : 5[3] -> 4[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO Connected all trees +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287868:1288720 [7] NCCL INFO comm 0x1fb84f10 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd54467debc3446a1 - Init COMPLETE +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO Connected all trees +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752205:2753088 [0] NCCL INFO comm 0x1e45d690 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x22aade7ce12590c4 - Init COMPLETE +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO Connected all trees +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO Connected all trees +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287866:1288714 [5] NCCL INFO comm 0x20894d10 rank 7 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x22aade7ce12590c4 - Init COMPLETE +ip-26-0-161-221:2752207:2753087 [2] NCCL INFO comm 0x1cb66dc0 rank 0 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd54467debc3446a1 - Init COMPLETE +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO Connected all trees +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287865:1288717 [4] NCCL INFO comm 0x2044f5f0 rank 6 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x22aade7ce12590c4 - Init COMPLETE +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO Connected all trees +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287867:1288718 [6] NCCL INFO comm 0x1f9acc50 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd54467debc3446a1 - Init COMPLETE +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO Connected all trees +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752209:2753091 [4] NCCL INFO comm 0x16517cf0 rank 2 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x22aade7ce12590c4 - Init COMPLETE +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO Connected all trees +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO Connected all trees +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO Connected all trees +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752206:2753093 [1] NCCL INFO comm 0x1cd1c7a0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x22aade7ce12590c4 - Init COMPLETE +ip-26-0-161-221:2752208:2753094 [3] NCCL INFO comm 0x1e55d850 rank 1 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd54467debc3446a1 - Init COMPLETE +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO Connected all trees +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO Connected all trees +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287862:1288721 [1] NCCL INFO comm 0x14b86cc0 rank 5 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x22aade7ce12590c4 - Init COMPLETE +ip-26-0-161-78:1287861:1288715 [0] NCCL INFO comm 0x15347dc0 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x22aade7ce12590c4 - Init COMPLETE +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO Connected all trees +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-78:1287863:1288719 [2] NCCL INFO comm 0x161d2c40 rank 4 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd54467debc3446a1 - Init COMPLETE +ip-26-0-161-78:1287864:1288716 [3] NCCL INFO comm 0x14428510 rank 5 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd54467debc3446a1 - Init COMPLETE +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO Connected all trees +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO Connected all trees +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO Connected all trees +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:2752210:2753089 [5] NCCL INFO comm 0x181847c0 rank 3 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x22aade7ce12590c4 - Init COMPLETE +ip-26-0-161-221:2752212:2753092 [7] NCCL INFO comm 0x15cc1e20 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd54467debc3446a1 - Init COMPLETE +ip-26-0-161-221:2752211:2753090 [6] NCCL INFO comm 0x160d6b60 rank 2 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd54467debc3446a1 - Init COMPLETE +[2024-12-27 17:30:23,290] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2752206 closing signal SIGTERM +[2024-12-27 17:30:23,290] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2752207 closing signal SIGTERM +[2024-12-27 17:30:23,290] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2752209 closing signal SIGTERM +[2024-12-27 17:30:23,290] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2752212 closing signal SIGTERM +[2024-12-27 17:30:23,311] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1287864 closing signal SIGTERM +[2024-12-27 17:30:23,311] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1287865 closing signal SIGTERM +[2024-12-27 17:30:23,475] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1287861) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:30:23 + host : ip-26-0-161-78.ec2.internal + rank : 9 (local_rank: 1) + exitcode : 1 (pid: 1287862) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_17:30:23 + host : ip-26-0-161-78.ec2.internal + rank : 10 (local_rank: 2) + exitcode : 1 (pid: 1287863) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_17:30:23 + host : ip-26-0-161-78.ec2.internal + rank : 13 (local_rank: 5) + exitcode : 1 (pid: 1287866) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-27_17:30:23 + host : ip-26-0-161-78.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 1287867) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-27_17:30:23 + host : ip-26-0-161-78.ec2.internal + rank : 15 (local_rank: 7) + exitcode : 1 (pid: 1287868) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:30:23 + host : ip-26-0-161-78.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 1287861) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-27 17:30:23,631] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2752205) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 17:30:23,670] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-221.ec2.internal_2752133_0' has failed to shutdown the rendezvous '13803238' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:30:23 + host : ip-26-0-161-221.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 2752208) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_17:30:23 + host : ip-26-0-161-221.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 2752210) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_17:30:23 + host : ip-26-0-161-221.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 2752211) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:30:23 + host : ip-26-0-161-221.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 2752205) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-161-78: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13803238.0 +srun: error: ip-26-0-161-221: task 1: Terminated +srun: Force Terminated StepId=13803238.0 diff --git a/logs/13803286-bench_80G_dp2_tp4_pp2_acc64_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13803286-bench_80G_dp2_tp4_pp2_acc64_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..8951e569d1995d77067c0510f2791dd7101167db --- /dev/null +++ b/logs/13803286-bench_80G_dp2_tp4_pp2_acc64_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,5112 @@ ++ '[' -z 13803286 ']' ++ unset FI_PROVIDER ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-165-[164,202]' ++ export 'NODELIST=ip-26-0-165-164 +ip-26-0-165-202' ++ NODELIST='ip-26-0-165-164 +ip-26-0-165-202' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-165-[164,202]' ++ export MASTER_NODE=ip-26-0-165-164 ++ MASTER_NODE=ip-26-0-165-164 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ export NCCL_P2P_LEVEL=LOC ++ NCCL_P2P_LEVEL=LOC ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-165-164' +Master node: ip-26-0-165-164 ++ echo 'All nodes: ip-26-0-165-164 +ip-26-0-165-202' +All nodes: ip-26-0-165-164 +ip-26-0-165-202 ++ echo 'World size: 16' +World size: 16 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13803286 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-165-164:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_80G_dp2_tp4_pp2_acc64_mbs2_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-27 17:27:54,510] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:27:54,536] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:27:54,510] torch.distributed.run: [WARNING] +[2024-12-27 17:27:54,510] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:27:54,510] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:27:54,510] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:27:54,536] torch.distributed.run: [WARNING] +[2024-12-27 17:27:54,536] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:27:54,536] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:27:54,536] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-165-164:1069368:1069368 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.165.164<0> +ip-26-0-165-164:1069368:1069368 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-164:1069368:1069368 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-164:1069368:1069368 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-165-202:1063487:1063487 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-202:1063485:1063485 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-164:1069369:1069369 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-164:1069369:1069369 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.165.164<0> +ip-26-0-165-164:1069372:1069372 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-164:1069375:1069375 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-164:1069369:1069369 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-164:1069369:1069369 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-202:1063485:1063485 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.165.202<0> +ip-26-0-165-202:1063487:1063487 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.165.202<0> +ip-26-0-165-202:1063486:1063486 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-202:1063481:1063481 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-202:1063487:1063487 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-202:1063485:1063485 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-202:1063487:1063487 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-202:1063485:1063485 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-164:1069372:1069372 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.165.164<0> +ip-26-0-165-164:1069375:1069375 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.165.164<0> +ip-26-0-165-164:1069372:1069372 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-164:1069372:1069372 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-164:1069373:1069373 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-164:1069375:1069375 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-164:1069375:1069375 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-164:1069373:1069373 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.165.164<0> +ip-26-0-165-164:1069374:1069374 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-202:1063486:1063486 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.165.202<0> +ip-26-0-165-202:1063481:1063481 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.165.202<0> +ip-26-0-165-164:1069373:1069373 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-164:1069373:1069373 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-164:1069371:1069371 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-164:1069370:1069370 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-202:1063481:1063481 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-202:1063481:1063481 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-202:1063486:1063486 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-202:1063486:1063486 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-164:1069374:1069374 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.165.164<0> +ip-26-0-165-164:1069374:1069374 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-164:1069374:1069374 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-202:1063483:1063483 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-164:1069370:1069370 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.165.164<0> +ip-26-0-165-164:1069371:1069371 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.165.164<0> +ip-26-0-165-164:1069370:1069370 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-164:1069370:1069370 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-164:1069371:1069371 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-164:1069371:1069371 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-202:1063483:1063483 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.165.202<0> +ip-26-0-165-202:1063483:1063483 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-202:1063483:1063483 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-202:1063484:1063484 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-202:1063480:1063480 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-202:1063482:1063482 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-165-202:1063484:1063484 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.165.202<0> +ip-26-0-165-202:1063484:1063484 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-202:1063484:1063484 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-202:1063480:1063480 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.165.202<0> +ip-26-0-165-202:1063482:1063482 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.165.202<0> +ip-26-0-165-202:1063480:1063480 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-202:1063480:1063480 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-202:1063482:1063482 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-165-202:1063482:1063482 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/205 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO comm 0x89f5f40 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO comm 0x8a38740 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO comm 0x8b97cd0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO comm 0x8eae590 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO comm 0x863c4a0 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO comm 0x8841cc0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO comm 0xa411740 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO comm 0x895f760 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO comm 0x99331d0 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO comm 0x90f50b0 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO comm 0x8acf3e0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO comm 0x8905b20 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO comm 0x9b08b00 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO comm 0xa0edcc0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO comm 0x9a3cf00 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO comm 0x974b940 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa4063e622096c6b - Init START +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 7/-1/-1->15->-1 [3] 7/-1/-1->15->-1 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Trees [0] 8/-1/-1->0->-1 [1] 8/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Trees [0] 4/12/-1->8->0 [1] 4/12/-1->8->0 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Connected all rings +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Connected all rings +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Connected all rings +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Connected all rings +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Connected all rings +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Connected all rings +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Connected all rings +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Connected all rings +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Connected all rings +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Connected all rings +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Connected all rings +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Connected all rings +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Connected all rings +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Connected all rings +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Connected all rings +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Connected all rings +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO Connected all trees +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO Connected all trees +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063487:1063639 [7] NCCL INFO comm 0x99331d0 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-164:1069368:1069530 [0] NCCL INFO comm 0x895f760 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO Connected all trees +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO Connected all trees +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO Connected all trees +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO Connected all trees +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO Connected all trees +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069370:1069537 [2] NCCL INFO comm 0x8b97cd0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-164:1069374:1069534 [6] NCCL INFO comm 0x8eae590 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO Connected all trees +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069369:1069531 [1] NCCL INFO comm 0x8a38740 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-202:1063482:1063645 [2] NCCL INFO comm 0x9b08b00 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-164:1069373:1069535 [5] NCCL INFO comm 0x89f5f40 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-202:1063486:1063641 [6] NCCL INFO comm 0x90f50b0 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO Connected all trees +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO Connected all trees +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063481:1063642 [1] NCCL INFO comm 0x9a3cf00 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-202:1063485:1063638 [5] NCCL INFO comm 0x8acf3e0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO Connected all trees +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO Connected all trees +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069371:1069536 [3] NCCL INFO comm 0xa411740 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO Connected all trees +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO Connected all trees +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069375:1069533 [7] NCCL INFO comm 0x8841cc0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO Connected all trees +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063483:1063640 [3] NCCL INFO comm 0x974b940 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-164:1069372:1069532 [4] NCCL INFO comm 0x863c4a0 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-202:1063480:1063643 [0] NCCL INFO comm 0x8905b20 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO Connected all trees +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063484:1063644 [4] NCCL INFO comm 0xa0edcc0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa4063e622096c6b - Init COMPLETE +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO comm 0x9b69930 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO comm 0x932c490 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO comm 0x8d06e00 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO comm 0x90e4f40 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO comm 0x8a780f0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO comm 0x8873620 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO comm 0xa6489c0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO comm 0x8c2c630 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO comm 0x8dcf4a0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO comm 0x8c6e9a0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO comm 0x8b99810 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO comm 0xa324cc0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO comm 0x8b3c560 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO comm 0x9983490 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO comm 0x9d3f9c0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO comm 0x9c739b0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb8e4ea16b77622d0 - Init START +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 7/-1/-1->15->-1 [3] 7/-1/-1->15->-1 +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Trees [0] 4/12/-1->8->0 [1] 4/12/-1->8->0 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Trees [0] 8/-1/-1->0->-1 [1] 8/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 01/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 02/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 03/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Connected all rings +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Connected all rings +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Connected all rings +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Connected all rings +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Connected all rings +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Connected all rings +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Connected all rings +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Connected all rings +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Connected all rings +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Connected all rings +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Connected all rings +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Connected all rings +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Connected all rings +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Connected all rings +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Connected all rings +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Connected all rings +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO Connected all trees +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO Connected all trees +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063487:1063876 [7] NCCL INFO comm 0x9b69930 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-164:1069368:1069760 [0] NCCL INFO comm 0x8b99810 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO Connected all trees +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO Connected all trees +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO Connected all trees +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069370:1069763 [2] NCCL INFO comm 0x8dcf4a0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO Connected all trees +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063482:1063873 [2] NCCL INFO comm 0x9d3f9c0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-164:1069369:1069764 [1] NCCL INFO comm 0x8c6e9a0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-202:1063486:1063869 [6] NCCL INFO comm 0x932c490 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO Connected all trees +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069374:1069762 [6] NCCL INFO comm 0x90e4f40 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO Connected all trees +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069373:1069761 [5] NCCL INFO comm 0x8c2c630 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO Connected all trees +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO Connected all trees +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069371:1069766 [3] NCCL INFO comm 0xa6489c0 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO Connected all trees +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO Connected all trees +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO Connected all trees +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063483:1063870 [3] NCCL INFO comm 0x9983490 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO Connected all trees +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO Connected all trees +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO Connected all trees +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063480:1063872 [0] NCCL INFO comm 0x8b3c560 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-202:1063481:1063874 [1] NCCL INFO comm 0x9c739b0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-164:1069375:1069767 [7] NCCL INFO comm 0x8a780f0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-164:1069372:1069765 [4] NCCL INFO comm 0x8873620 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-202:1063485:1063871 [5] NCCL INFO comm 0x8d06e00 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +ip-26-0-165-202:1063484:1063875 [4] NCCL INFO comm 0xa324cc0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb8e4ea16b77622d0 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO comm 0x9b7dc00 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x358d75a7dccf288d - Init START +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO comm 0x8b52440 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x358d75a7dccf288d - Init START +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO comm 0x9340b10 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x358d75a7dccf288d - Init START +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO comm 0x8d1ab50 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x358d75a7dccf288d - Init START +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO comm 0xa338aa0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x358d75a7dccf288d - Init START +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO comm 0x9c87ea0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x358d75a7dccf288d - Init START +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO comm 0x9996d90 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x358d75a7dccf288d - Init START +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO comm 0x9d53990 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x358d75a7dccf288d - Init START +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO comm 0x90f9880 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8a8631a7ba4df291 - Init START +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO comm 0x8c40a40 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8a8631a7ba4df291 - Init START +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO comm 0x8887940 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8a8631a7ba4df291 - Init START +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO comm 0x8de3250 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8a8631a7ba4df291 - Init START +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO comm 0x8bae300 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8a8631a7ba4df291 - Init START +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO comm 0x8c82910 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8a8631a7ba4df291 - Init START +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO comm 0x8a8c5d0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8a8631a7ba4df291 - Init START +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO comm 0xa65cc40 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8a8631a7ba4df291 - Init START +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Connected all rings +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Connected all rings +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Connected all rings +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Connected all rings +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Connected all rings +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Connected all rings +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Connected all rings +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Connected all rings +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO Connected all trees +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063487:1063907 [7] NCCL INFO comm 0x9b7dc00 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x358d75a7dccf288d - Init COMPLETE +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO Connected all trees +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063480:1063902 [0] NCCL INFO comm 0x8b52440 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x358d75a7dccf288d - Init COMPLETE +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Connected all rings +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Connected all rings +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Connected all rings +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Connected all rings +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Connected all rings +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Connected all rings +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Connected all rings +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Connected all rings +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO Connected all trees +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063905 [1] NCCL INFO comm 0x9c87ea0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x358d75a7dccf288d - Init COMPLETE +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO Connected all trees +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO Connected all trees +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063486:1063906 [6] NCCL INFO comm 0x9340b10 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x358d75a7dccf288d - Init COMPLETE +ip-26-0-165-202:1063482:1063903 [2] NCCL INFO comm 0x9d53990 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x358d75a7dccf288d - Init COMPLETE +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO Connected all trees +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063483:1063909 [3] NCCL INFO comm 0x9996d90 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x358d75a7dccf288d - Init COMPLETE +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO Connected all trees +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO Connected all trees +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063485:1063908 [5] NCCL INFO comm 0x8d1ab50 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x358d75a7dccf288d - Init COMPLETE +ip-26-0-165-202:1063484:1063904 [4] NCCL INFO comm 0xa338aa0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x358d75a7dccf288d - Init COMPLETE +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO Connected all trees +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069375:1069802 [7] NCCL INFO comm 0x8a8c5d0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8a8631a7ba4df291 - Init COMPLETE +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO Connected all trees +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069368:1069795 [0] NCCL INFO comm 0x8bae300 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8a8631a7ba4df291 - Init COMPLETE +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO Connected all trees +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO Connected all trees +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069369:1069798 [1] NCCL INFO comm 0x8c82910 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8a8631a7ba4df291 - Init COMPLETE +ip-26-0-165-164:1069374:1069796 [6] NCCL INFO comm 0x90f9880 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8a8631a7ba4df291 - Init COMPLETE +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO Connected all trees +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO Connected all trees +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069371:1069801 [3] NCCL INFO comm 0xa65cc40 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8a8631a7ba4df291 - Init COMPLETE +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO Connected all trees +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069370:1069799 [2] NCCL INFO comm 0x8de3250 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8a8631a7ba4df291 - Init COMPLETE +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO Connected all trees +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069373:1069797 [5] NCCL INFO comm 0x8c40a40 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8a8631a7ba4df291 - Init COMPLETE +ip-26-0-165-164:1069372:1069800 [4] NCCL INFO comm 0x8887940 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8a8631a7ba4df291 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO comm 0xaba5260 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xba38a8ca6c9c6739 - Init START +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO comm 0xad031b0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xba38a8ca6c9c6739 - Init START +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO comm 0xaad1420 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xba38a8ca6c9c6739 - Init START +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO comm 0xc57d660 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xba38a8ca6c9c6739 - Init START +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO comm 0xab60060 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5792bc05c1a912eb - Init START +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO comm 0xa7a89e0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5792bc05c1a912eb - Init START +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO comm 0xb018ba0 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5792bc05c1a912eb - Init START +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO comm 0xa9adce0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5792bc05c1a912eb - Init START +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO comm 0xa8ae720 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xbdf2a61e3d13eff2 - Init START +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO comm 0xb6f0c60 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbdf2a61e3d13eff2 - Init START +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO comm 0xbaad5e0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xbdf2a61e3d13eff2 - Init START +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO comm 0xb9e2e60 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbdf2a61e3d13eff2 - Init START +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO comm 0xc093160 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb2cf0edfb9deb293 - Init START +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO comm 0xaa741e0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb2cf0edfb9deb293 - Init START +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO comm 0xb8d7720 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb2cf0edfb9deb293 - Init START +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO comm 0xb09ae70 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb2cf0edfb9deb293 - Init START +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Connected all rings +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Connected all rings +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Connected all rings +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Connected all rings +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Connected all rings +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Connected all rings +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Connected all rings +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Connected all rings +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Connected all rings +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Connected all rings +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Connected all rings +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Connected all rings +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Connected all rings +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Connected all rings +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Connected all rings +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Connected all rings +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO Connected all trees +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO Connected all trees +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069371:1069841 [3] NCCL INFO comm 0xc57d660 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xba38a8ca6c9c6739 - Init COMPLETE +ip-26-0-165-164:1069372:1069833 [4] NCCL INFO comm 0xa7a89e0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5792bc05c1a912eb - Init COMPLETE +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO Connected all trees +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069375:1069836 [7] NCCL INFO comm 0xa9adce0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5792bc05c1a912eb - Init COMPLETE +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO Connected all trees +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069368:1069835 [0] NCCL INFO comm 0xaad1420 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xba38a8ca6c9c6739 - Init COMPLETE +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO Connected all trees +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO Connected all trees +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO Connected all trees +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO Connected all trees +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO Connected all trees +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO Connected all trees +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069374:1069837 [6] NCCL INFO comm 0xb018ba0 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5792bc05c1a912eb - Init COMPLETE +ip-26-0-165-164:1069369:1069840 [1] NCCL INFO comm 0xaba5260 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xba38a8ca6c9c6739 - Init COMPLETE +ip-26-0-165-164:1069370:1069839 [2] NCCL INFO comm 0xad031b0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xba38a8ca6c9c6739 - Init COMPLETE +ip-26-0-165-202:1063483:1063945 [3] NCCL INFO comm 0xb6f0c60 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbdf2a61e3d13eff2 - Init COMPLETE +ip-26-0-165-164:1069373:1069838 [5] NCCL INFO comm 0xab60060 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5792bc05c1a912eb - Init COMPLETE +ip-26-0-165-202:1063487:1063947 [7] NCCL INFO comm 0xb8d7720 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb2cf0edfb9deb293 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO comm 0xb703920 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5010b588e9e7d9cf - Init START +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO comm 0xc590320 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5010b588e9e7d9cf - Init START +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO comm 0xa9c09a0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x19b67ddc6c29686c - Init START +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO comm 0xb8ea3e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x19b67ddc6c29686c - Init START +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO Connected all trees +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063480:1063940 [0] NCCL INFO comm 0xa8ae720 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xbdf2a61e3d13eff2 - Init COMPLETE +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO comm 0xa8c13e0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2f0e70142fd8cada - Init START +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO comm 0xaae40e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2f0e70142fd8cada - Init START +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO Connected all trees +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063484:1063942 [4] NCCL INFO comm 0xc093160 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb2cf0edfb9deb293 - Init COMPLETE +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO Connected all trees +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO Connected all trees +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO comm 0xc0a5e20 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x47d8b262c5e1f51 - Init START +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO comm 0xa7bb6a0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x47d8b262c5e1f51 - Init START +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO Connected all trees +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063482:1063943 [2] NCCL INFO comm 0xbaad5e0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xbdf2a61e3d13eff2 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-165-202:1063481:1063944 [1] NCCL INFO comm 0xb9e2e60 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbdf2a61e3d13eff2 - Init COMPLETE +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063486:1063946 [6] NCCL INFO comm 0xb09ae70 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb2cf0edfb9deb293 - Init COMPLETE +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO comm 0xbac02a0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb4cc3513948e0adc - Init START +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO comm 0xad15e70 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb4cc3513948e0adc - Init START +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO comm 0xb9f5b20 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa36a49b0843386aa - Init START +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO comm 0xabb7f20 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa36a49b0843386aa - Init START +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO comm 0xb02b860 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x636567cc99d5b284 - Init START +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO comm 0xb0adb30 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x636567cc99d5b284 - Init START +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO Connected all trees +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063485:1063948 [5] NCCL INFO comm 0xaa741e0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb2cf0edfb9deb293 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO comm 0xaa86ea0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb29951fef5c25177 - Init START +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO comm 0xab72d20 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb29951fef5c25177 - Init START +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Connected all rings +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO Connected all trees +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069371:1069863 [3] NCCL INFO comm 0xc590320 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5010b588e9e7d9cf - Init COMPLETE +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Connected all rings +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO Connected all trees +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Connected all rings +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO Connected all trees +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Connected all rings +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO Connected all trees +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Connected all rings +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO Connected all trees +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Connected all rings +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO Connected all trees +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069368:1069867 [0] NCCL INFO comm 0xaae40e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2f0e70142fd8cada - Init COMPLETE +ip-26-0-165-202:1063483:1063969 [3] NCCL INFO comm 0xb703920 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5010b588e9e7d9cf - Init COMPLETE +ip-26-0-165-164:1069375:1069865 [7] NCCL INFO comm 0xa9c09a0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x19b67ddc6c29686c - Init COMPLETE +ip-26-0-165-202:1063480:1063971 [0] NCCL INFO comm 0xa8c13e0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2f0e70142fd8cada - Init COMPLETE +ip-26-0-165-202:1063487:1063970 [7] NCCL INFO comm 0xb8ea3e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x19b67ddc6c29686c - Init COMPLETE +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Connected all rings +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO Connected all trees +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Connected all rings +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO Connected all trees +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Connected all rings +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO Connected all trees +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Connected all rings +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO Connected all trees +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069372:1069875 [4] NCCL INFO comm 0xa7bb6a0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x47d8b262c5e1f51 - Init COMPLETE +ip-26-0-165-164:1069370:1069877 [2] NCCL INFO comm 0xad15e70 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb4cc3513948e0adc - Init COMPLETE +ip-26-0-165-164:1069369:1069879 [1] NCCL INFO comm 0xabb7f20 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa36a49b0843386aa - Init COMPLETE +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Connected all rings +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO Connected all trees +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063482:1063978 [2] NCCL INFO comm 0xbac02a0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb4cc3513948e0adc - Init COMPLETE +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Connected all rings +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO Connected all trees +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Connected all rings +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO Connected all trees +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063481:1063979 [1] NCCL INFO comm 0xb9f5b20 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa36a49b0843386aa - Init COMPLETE +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Connected all rings +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO Connected all trees +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Connected all rings +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO Connected all trees +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069374:1069881 [6] NCCL INFO comm 0xb02b860 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x636567cc99d5b284 - Init COMPLETE +ip-26-0-165-202:1063484:1063977 [4] NCCL INFO comm 0xc0a5e20 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x47d8b262c5e1f51 - Init COMPLETE +ip-26-0-165-164:1069373:1069883 [5] NCCL INFO comm 0xab72d20 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb29951fef5c25177 - Init COMPLETE +ip-26-0-165-202:1063486:1063980 [6] NCCL INFO comm 0xb0adb30 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x636567cc99d5b284 - Init COMPLETE +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Connected all rings +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO Connected all trees +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063485:1063982 [5] NCCL INFO comm 0xaa86ea0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb29951fef5c25177 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO comm 0xad28b30 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x16bc2a9cea99d6de - Init START +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO comm 0xb03ed30 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x16bc2a9cea99d6de - Init START +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO comm 0xb0ce6d0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x12ebba9d2137639a - Init START +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO comm 0xbae1570 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x12ebba9d2137639a - Init START +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO comm 0xa7ce360 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5d508ded4288335d - Init START +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO comm 0xaaf6da0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5d508ded4288335d - Init START +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO comm 0xa8e2830 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x16b16d27e3abf5f1 - Init START +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO comm 0xc0c7730 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x16b16d27e3abf5f1 - Init START +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO comm 0xc5a37f0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xda317c6b8e5e058e - Init START +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO comm 0xa9d3660 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xda317c6b8e5e058e - Init START +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Using network Libfabric +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO comm 0xab859e0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeee7bfa9f43b1db0 - Init START +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO comm 0xabcabe0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeee7bfa9f43b1db0 - Init START +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +NCCL version 2.18.5+cuda12.2 +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO comm 0xaaa8c00 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcc0f69588f01226f - Init START +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO comm 0xba16f10 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xcc0f69588f01226f - Init START +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Using network Libfabric +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO comm 0xb90b310 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3f1fcab0c2dc21a - Init START +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO comm 0xb724640 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3f1fcab0c2dc21a - Init START +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Connected all rings +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO Connected all trees +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Connected all rings +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO Connected all trees +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Connected all rings +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO Connected all trees +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069374:1069897 [6] NCCL INFO comm 0xb03ed30 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x16bc2a9cea99d6de - Init COMPLETE +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Connected all rings +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO Connected all trees +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Connected all rings +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO Connected all trees +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Connected all rings +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO Connected all trees +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069370:1069896 [2] NCCL INFO comm 0xad28b30 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x16bc2a9cea99d6de - Init COMPLETE +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Connected all rings +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO Connected all trees +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Connected all rings +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO Connected all trees +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063481:1064002 [1] NCCL INFO comm 0xba16f10 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xcc0f69588f01226f - Init COMPLETE +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Connected all rings +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO Connected all trees +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063482:1063996 [2] NCCL INFO comm 0xbae1570 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x12ebba9d2137639a - Init COMPLETE +ip-26-0-165-202:1063486:1063997 [6] NCCL INFO comm 0xb0ce6d0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x12ebba9d2137639a - Init COMPLETE +ip-26-0-165-202:1063485:1064003 [5] NCCL INFO comm 0xaaa8c00 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcc0f69588f01226f - Init COMPLETE +ip-26-0-165-202:1063484:1064000 [4] NCCL INFO comm 0xc0c7730 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x16b16d27e3abf5f1 - Init COMPLETE +ip-26-0-165-202:1063480:1063999 [0] NCCL INFO comm 0xa8e2830 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x16b16d27e3abf5f1 - Init COMPLETE +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Connected all rings +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO Connected all trees +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Connected all rings +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO Connected all trees +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069371:1069902 [3] NCCL INFO comm 0xc5a37f0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xda317c6b8e5e058e - Init COMPLETE +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Connected all rings +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO Connected all trees +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Connected all rings +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO Connected all trees +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069375:1069903 [7] NCCL INFO comm 0xa9d3660 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xda317c6b8e5e058e - Init COMPLETE +ip-26-0-165-202:1063487:1064006 [7] NCCL INFO comm 0xb90b310 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3f1fcab0c2dc21a - Init COMPLETE +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Connected all rings +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO Connected all trees +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-202:1063483:1064005 [3] NCCL INFO comm 0xb724640 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3f1fcab0c2dc21a - Init COMPLETE +ip-26-0-165-164:1069368:1069899 [0] NCCL INFO comm 0xaaf6da0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5d508ded4288335d - Init COMPLETE +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Connected all rings +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO Connected all trees +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Connected all rings +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO Connected all trees +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-165-164:1069369:1069905 [1] NCCL INFO comm 0xabcabe0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeee7bfa9f43b1db0 - Init COMPLETE +ip-26-0-165-164:1069372:1069900 [4] NCCL INFO comm 0xa7ce360 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5d508ded4288335d - Init COMPLETE +ip-26-0-165-164:1069373:1069906 [5] NCCL INFO comm 0xab859e0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeee7bfa9f43b1db0 - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader)trainer.train(dataloader) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0]return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self._call_impl(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return forward_call(*args, **kwargs) +return self.act(gate_states) * up_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 65.94 MiB is free. Including non-PyTorch memory, this process has 79.25 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 322.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.act(gate_states) * up_states +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 1.94 MiB is free. Including non-PyTorch memory, this process has 79.32 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 306.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.act(gate_states) * up_states +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 65.94 MiB is free. Including non-PyTorch memory, this process has 79.25 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 322.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 65.94 MiB is free. Including non-PyTorch memory, this process has 79.25 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 322.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) +return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.act(gate_states) * up_states +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 65.94 MiB is free. Including non-PyTorch memory, this process has 79.25 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 322.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self.act(gate_states) * up_states +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 65.94 MiB is free. Including non-PyTorch memory, this process has 79.25 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 322.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return self.act(gate_states) * up_states +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 1.94 MiB is free. Including non-PyTorch memory, this process has 79.32 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 306.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 466, in forward + out = F.linear(tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 128.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 113.94 MiB is free. Including non-PyTorch memory, this process has 79.21 GiB memory in use. Of the allocated memory 77.26 GiB is allocated by PyTorch, and 322.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[E TraceUtils.h:35] Store is down while updating #4 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #4 with key NCCL_2_trace_start +[E TraceUtils.h:35] Store is down while updating #5 with key NCCL_3_trace_start +[E TraceUtils.h:35] Store is down while updating #234 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #4 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #4 with key NCCL_3_trace_start +[E TraceUtils.h:35] Store is down while updating #4 with key NCCL_2_trace_start +[E TraceUtils.h:35] Store is down while updating #4 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #233 with key NCCL_0_trace_end +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=2|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=3|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=2|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=2|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=3|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-165-202]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-165-202]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-165-202]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=3|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=2|ip-26-0-165-202]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-165-202]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=2|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=3|ip-26-0-165-202]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=2|ip-26-0-165-202]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=3|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=3|ip-26-0-165-202]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=2|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=2|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=3|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=2|ip-26-0-165-202]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-165-202]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=2|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-165-202]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=3|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-165-202]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-165-202]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=3|ip-26-0-165-202]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=2|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=2|ip-26-0-165-202]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=3|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=3|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=3|ip-26-0-165-202]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=2|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=2|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=3|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=2|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=2|ip-26-0-165-202]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-165-202]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-165-202]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-165-202]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=3|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=2|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-165-202]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=3|ip-26-0-165-202]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=2|ip-26-0-165-202]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=3|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=3|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=3|ip-26-0-165-202]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=2|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=2|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=3|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=2|ip-26-0-165-202]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=2|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-165-202]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-165-202]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=3|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-165-202]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=2|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=3|ip-26-0-165-202]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=2|ip-26-0-165-202]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-165-202]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=3|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=3|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=3|ip-26-0-165-202]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=2|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=2|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=2|ip-26-0-165-202]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=3|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=2|ip-26-0-165-202]: Max retries reached, giving up on communication +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=2|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=0|ip-26-0-165-202]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=3|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=1|ip-26-0-165-202]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=0|ip-26-0-165-202]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=0|ip-26-0-165-202]: Max retries reached, giving up on communication +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=2|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=3|ip-26-0-165-202]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=1|ip-26-0-165-202]: Max retries reached, giving up on communication +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=0|ip-26-0-165-202]: Max retries reached, giving up on communication +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=2|ip-26-0-165-202]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=3|ip-26-0-165-202]: Max retries reached, giving up on communication +12/27/2024 17:28:42 [WARNING|DP=0|PP=1|TP=1|ip-26-0-165-202]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=2|ip-26-0-165-202]: Max retries reached, giving up on communication +12/27/2024 17:28:42 [ERROR|DP=0|PP=1|TP=1|ip-26-0-165-202]: Max retries reached, giving up on communication +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=3|ip-26-0-165-202]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=3|ip-26-0-165-202]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:28:42 [WARNING|DP=1|PP=1|TP=3|ip-26-0-165-202]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:28:42 [ERROR|DP=1|PP=1|TP=3|ip-26-0-165-202]: Max retries reached, giving up on communication +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +trainer.train(dataloader) + trainer.train(dataloader) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +return f(*args, **kwargs) + return f(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)outputs = self.pipeline_engine.train_batch_iter( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + output = model(**micro_batch) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + output = model(**micro_batch) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + sharded_logits = self.model(output = model(**micro_batch) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + activation = pipeline_state.activations_buffer.popleft() File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + +IndexError: pop from an empty deque + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states)hidden_encoder_states = encoder_block(**hidden_encoder_states) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer +new_kwargs[name] = recv_from_pipeline_state_buffer( +new_kwargs[name] = recv_from_pipeline_state_buffer( File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + activation = pipeline_state.activations_buffer.popleft()activation = pipeline_state.activations_buffer.popleft() + +IndexErrorIndexError: pop from an empty deque: +pop from an empty deque + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque +[2024-12-27 17:29:05,787] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1069369 closing signal SIGTERM +[2024-12-27 17:29:05,787] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1069373 closing signal SIGTERM +[2024-12-27 17:29:05,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1063481 closing signal SIGTERM +[2024-12-27 17:29:05,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1063484 closing signal SIGTERM +[2024-12-27 17:29:05,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1063485 closing signal SIGTERM +[2024-12-27 17:29:05,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1063486 closing signal SIGTERM +[2024-12-27 17:29:05,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1063487 closing signal SIGTERM +[2024-12-27 17:29:06,352] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1069368) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 17:29:06,407] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_69rqmqs_/13803286_81ukxvaq/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:28:41 + host : ip-26-0-165-164.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 1069370) + error_file: /tmp/torchelastic_69rqmqs_/13803286_81ukxvaq/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 65.94 MiB is free. Including non-PyTorch memory, this process has 79.25 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 322.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[2]: + time : 2024-12-27_17:28:41 + host : ip-26-0-165-164.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 1069371) + error_file: /tmp/torchelastic_69rqmqs_/13803286_81ukxvaq/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 1.94 MiB is free. Including non-PyTorch memory, this process has 79.32 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 306.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[3]: + time : 2024-12-27_17:28:41 + host : ip-26-0-165-164.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 1069372) + error_file: /tmp/torchelastic_69rqmqs_/13803286_81ukxvaq/attempt_0/4/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 1.94 MiB is free. Including non-PyTorch memory, this process has 79.32 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 306.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[4]: + time : 2024-12-27_17:28:41 + host : ip-26-0-165-164.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 1069374) + error_file: /tmp/torchelastic_69rqmqs_/13803286_81ukxvaq/attempt_0/6/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 65.94 MiB is free. Including non-PyTorch memory, this process has 79.25 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 322.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[5]: + time : 2024-12-27_17:28:41 + host : ip-26-0-165-164.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 1069375) + error_file: /tmp/torchelastic_69rqmqs_/13803286_81ukxvaq/attempt_0/7/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 65.94 MiB is free. Including non-PyTorch memory, this process has 79.25 GiB memory in use. Of the allocated memory 77.15 GiB is allocated by PyTorch, and 322.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:28:41 + host : ip-26-0-165-164.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 1069368) + error_file: /tmp/torchelastic_69rqmqs_/13803286_81ukxvaq/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 466, in forward + out = F.linear(tensor, weight, bias) + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 128.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 113.94 MiB is free. Including non-PyTorch memory, this process has 79.21 GiB memory in use. Of the allocated memory 77.26 GiB is allocated by PyTorch, and 322.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +============================================================ +[2024-12-27 17:29:06,468] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1063480) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 17:29:06,505] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-202.ec2.internal_1063410_0' has failed to shutdown the rendezvous '13803286' due to an error of type RendezvousConnectionError. +[2024-12-27 17:29:06,505] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_ix9adynw/13803286_1o80vfwa/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:28:42 + host : ip-26-0-165-202.ec2.internal + rank : 10 (local_rank: 2) + exitcode : 1 (pid: 1063482) + error_file: /tmp/torchelastic_ix9adynw/13803286_1o80vfwa/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +[2]: + time : 2024-12-27_17:28:42 + host : ip-26-0-165-202.ec2.internal + rank : 11 (local_rank: 3) + exitcode : 1 (pid: 1063483) + error_file: /tmp/torchelastic_ix9adynw/13803286_1o80vfwa/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:28:42 + host : ip-26-0-165-202.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 1063480) + error_file: /tmp/torchelastic_ix9adynw/13803286_1o80vfwa/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +============================================================ +srun: error: ip-26-0-165-164: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13803286.0 +srun: error: ip-26-0-165-202: task 1: Terminated +srun: Force Terminated StepId=13803286.0 diff --git a/logs/13803408-bench_80G_dp2_tp1_pp16_acc16_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13803408-bench_80G_dp2_tp1_pp16_acc16_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..0bcc7de7f2a509f3cf038fb0ed23a63d084b647c --- /dev/null +++ b/logs/13803408-bench_80G_dp2_tp1_pp16_acc16_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,10054 @@ ++ '[' -z 13803408 ']' ++ unset FI_PROVIDER ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames ip-26-0-163-43,ip-26-0-164-187,ip-26-0-166-244,ip-26-0-167-9 ++ export 'NODELIST=ip-26-0-163-43 +ip-26-0-164-187 +ip-26-0-166-244 +ip-26-0-167-9' ++ NODELIST='ip-26-0-163-43 +ip-26-0-164-187 +ip-26-0-166-244 +ip-26-0-167-9' +++ scontrol show hostnames ip-26-0-163-43,ip-26-0-164-187,ip-26-0-166-244,ip-26-0-167-9 +++ head -n1 ++ export MASTER_NODE=ip-26-0-163-43 ++ MASTER_NODE=ip-26-0-163-43 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ export NCCL_P2P_LEVEL=LOC ++ NCCL_P2P_LEVEL=LOC ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-43 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 52-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 52-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 52-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 52-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-163-43' +Master node: ip-26-0-163-43 ++ echo 'All nodes: ip-26-0-163-43 +ip-26-0-164-187 +ip-26-0-166-244 +ip-26-0-167-9' +All nodes: ip-26-0-163-43 +ip-26-0-164-187 +ip-26-0-166-244 +ip-26-0-167-9 ++ echo 'World size: 32' +World size: 32 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13803408 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-163-43:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_80G_dp2_tp1_pp16_acc16_mbs8_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-27 17:34:47,428] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:34:47,428] torch.distributed.run: [WARNING] +[2024-12-27 17:34:47,428] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:34:47,428] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:34:47,428] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:34:47,466] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:34:47,496] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:34:47,466] torch.distributed.run: [WARNING] +[2024-12-27 17:34:47,466] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:34:47,466] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:34:47,466] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:34:47,496] torch.distributed.run: [WARNING] +[2024-12-27 17:34:47,496] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:34:47,496] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:34:47,496] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:34:47,687] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:34:47,687] torch.distributed.run: [WARNING] +[2024-12-27 17:34:47,687] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:34:47,687] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:34:47,687] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-163-43:2656306:2656306 [0] NCCL INFO Bootstrap : Using enp74s0:26.0.164.42<0> +ip-26-0-163-43:2656306:2656306 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-43:2656306:2656306 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-43:2656306:2656306 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-167-9:1070842:1070842 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1070837:1070837 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1070840:1070840 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110731:2110731 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110733:2110733 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1070842:1070842 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-167-9:1070842:1070842 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1070842:1070842 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1070837:1070837 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-167-9:1070840:1070840 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-164-187:2110737:2110737 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1070837:1070837 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1070837:1070837 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-164-187:2110736:2110736 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1070840:1070840 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1070840:1070840 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074710:1074710 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-43:2656307:2656307 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1074711:1074711 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110731:2110731 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.164.187<0> +ip-26-0-163-43:2656308:2656308 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110733:2110733 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.164.187<0> +ip-26-0-164-187:2110731:2110731 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-187:2110731:2110731 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074708:1074708 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-43:2656310:2656310 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1074707:1074707 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110733:2110733 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-187:2110733:2110733 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-43:2656309:2656309 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110737:2110737 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.164.187<0> +ip-26-0-164-187:2110736:2110736 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.164.187<0> +ip-26-0-166-244:1074709:1074709 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-43:2656313:2656313 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-43:2656312:2656312 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110737:2110737 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-43:2656311:2656311 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110737:2110737 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-164-187:2110736:2110736 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-187:2110736:2110736 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074705:1074705 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110730:2110730 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1074706:1074706 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1074712:1074712 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1070844:1070844 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110732:2110732 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110735:2110735 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1070838:1070838 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110734:2110734 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1070839:1070839 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1070844:1070844 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-167-9:1070838:1070838 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-164-187:2110730:2110730 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.164.187<0> +ip-26-0-167-9:1070838:1070838 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1070844:1070844 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1070844:1070844 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1070838:1070838 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1070843:1070843 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1070841:1070841 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-187:2110730:2110730 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-187:2110730:2110730 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-164-187:2110732:2110732 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.164.187<0> +ip-26-0-164-187:2110735:2110735 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.164.187<0> +ip-26-0-164-187:2110734:2110734 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.164.187<0> +ip-26-0-163-43:2656307:2656307 [1] NCCL INFO Bootstrap : Using enp74s0:26.0.164.42<0> +ip-26-0-163-43:2656308:2656308 [2] NCCL INFO Bootstrap : Using enp74s0:26.0.164.42<0> +ip-26-0-164-187:2110734:2110734 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-187:2110734:2110734 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1070839:1070839 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-164-187:2110732:2110732 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-187:2110735:2110735 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-187:2110735:2110735 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-164-187:2110732:2110732 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074711:1074711 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-163-43:2656308:2656308 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-43:2656308:2656308 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074710:1074710 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-163-43:2656307:2656307 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-43:2656307:2656307 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-43:2656310:2656310 [4] NCCL INFO Bootstrap : Using enp74s0:26.0.164.42<0> +ip-26-0-167-9:1070839:1070839 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1070839:1070839 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-43:2656309:2656309 [3] NCCL INFO Bootstrap : Using enp74s0:26.0.164.42<0> +ip-26-0-163-43:2656311:2656311 [5] NCCL INFO Bootstrap : Using enp74s0:26.0.164.42<0> +ip-26-0-163-43:2656312:2656312 [6] NCCL INFO Bootstrap : Using enp74s0:26.0.164.42<0> +ip-26-0-163-43:2656313:2656313 [7] NCCL INFO Bootstrap : Using enp74s0:26.0.164.42<0> +ip-26-0-166-244:1074711:1074711 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1074711:1074711 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074707:1074707 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-167-9:1070843:1070843 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-166-244:1074710:1074710 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1074710:1074710 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1070841:1070841 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-166-244:1074708:1074708 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-163-43:2656310:2656310 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-43:2656310:2656310 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-43:2656309:2656309 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-43:2656309:2656309 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074709:1074709 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-167-9:1070843:1070843 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1070843:1070843 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074705:1074705 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-166-244:1074706:1074706 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-166-244:1074712:1074712 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-163-43:2656312:2656312 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-43:2656312:2656312 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074707:1074707 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1074707:1074707 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-43:2656311:2656311 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-43:2656311:2656311 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-43:2656313:2656313 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-43:2656313:2656313 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1070841:1070841 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1070841:1070841 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074708:1074708 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1074708:1074708 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074709:1074709 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1074709:1074709 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074706:1074706 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1074706:1074706 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074712:1074712 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1074712:1074712 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1074705:1074705 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1074705:1074705 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/253 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO comm 0x94f4760 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO comm 0x8a45520 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO comm 0x9f02020 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO comm 0x8e43090 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO comm 0x90f38f0 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO comm 0x84fbbb0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO comm 0x8b445e0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO comm 0x944f240 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO comm 0x9b952d0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO comm 0x94cb9c0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO comm 0x87977f0 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO comm 0x8b986d0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO comm 0x89642c0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO comm 0x967b580 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO comm 0x9d9a4e0 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO comm 0xa1856f0 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO comm 0x8d9d6c0 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO comm 0x8a3bb60 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO comm 0xa0f9020 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO comm 0x94b8290 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO comm 0x9677400 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO comm 0x9baead0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO comm 0x9978890 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO comm 0x8ce5c30 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO comm 0x8529040 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO comm 0x8592b00 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO comm 0x90499f0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO comm 0x9abb0b0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO comm 0x8ed28d0 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO comm 0x84f9810 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO comm 0x99101c0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO comm 0x9d62b50 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6c8cdb6f08cb89f6 - Init START +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Trees [0] 16/-1/-1->0->-1 [1] 16/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 15/-1/-1->31->-1 [3] 15/-1/-1->31->-1 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Trees [0] 8/24/-1->16->0 [1] 8/24/-1->16->0 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Connected all rings +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Connected all rings +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Connected all rings +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Connected all rings +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Connected all rings +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Connected all rings +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Connected all rings +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Connected all rings +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Connected all rings +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Connected all rings +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Connected all rings +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Connected all rings +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Connected all rings +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070844:1071006 [7] NCCL INFO comm 0x94f4760 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO Connected all trees +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656306:2656491 [0] NCCL INFO comm 0x8e43090 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO Connected all trees +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO Connected all trees +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO Connected all trees +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO Connected all trees +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO Connected all trees +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110736:2110895 [6] NCCL INFO comm 0x8529040 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-163-43:2656312:2656497 [6] NCCL INFO comm 0x8b445e0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-164-187:2110731:2110896 [1] NCCL INFO comm 0x8ce5c30 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656308:2656493 [2] NCCL INFO comm 0x9f02020 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-164-187:2110732:2110900 [2] NCCL INFO comm 0x9baead0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO Connected all trees +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO Connected all trees +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO Connected all trees +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074711:1074874 [6] NCCL INFO comm 0xa0f9020 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-166-244:1074707:1074877 [2] NCCL INFO comm 0x8ed28d0 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-166-244:1074710:1074879 [5] NCCL INFO comm 0x99101c0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-167-9:1070839:1071008 [2] NCCL INFO comm 0x8d9d6c0 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-167-9:1070843:1071009 [6] NCCL INFO comm 0x944f240 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-166-244:1074706:1074875 [1] NCCL INFO comm 0x84f9810 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-164-187:2110735:2110899 [5] NCCL INFO comm 0xa1856f0 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070838:1071010 [1] NCCL INFO comm 0x8a3bb60 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-163-43:2656307:2656494 [1] NCCL INFO comm 0x8a45520 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-163-43:2656311:2656500 [5] NCCL INFO comm 0x94cb9c0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-167-9:1070842:1071005 [5] NCCL INFO comm 0x87977f0 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO Connected all trees +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO Connected all trees +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO Connected all trees +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656313:2656498 [7] NCCL INFO comm 0x9b952d0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO Connected all trees +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO Connected all trees +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO Connected all trees +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110737:2110901 [7] NCCL INFO comm 0x9d9a4e0 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110733:2110897 [3] NCCL INFO comm 0x9677400 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-166-244:1074708:1074878 [3] NCCL INFO comm 0x9d62b50 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-167-9:1070840:1071004 [3] NCCL INFO comm 0x8b986d0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-163-43:2656309:2656499 [3] NCCL INFO comm 0x84fbbb0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO Connected all trees +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656310:2656496 [4] NCCL INFO comm 0x90f38f0 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110734:2110902 [4] NCCL INFO comm 0x9978890 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-166-244:1074712:1074872 [7] NCCL INFO comm 0x90499f0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070841:1071011 [4] NCCL INFO comm 0x89642c0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-164-187:2110730:2110898 [0] NCCL INFO comm 0x967b580 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-166-244:1074709:1074876 [4] NCCL INFO comm 0x8592b00 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070837:1071007 [0] NCCL INFO comm 0x94b8290 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-166-244:1074705:1074873 [0] NCCL INFO comm 0x9abb0b0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6c8cdb6f08cb89f6 - Init COMPLETE +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO comm 0x872f830 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc666fb977eaae014 - Init START +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO comm 0x8c78740 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc666fb977eaae014 - Init START +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO comm 0x9702210 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc666fb977eaae014 - Init START +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO comm 0xa137c00 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc666fb977eaae014 - Init START +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO comm 0x9329580 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc666fb977eaae014 - Init START +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO comm 0x9087740 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc666fb977eaae014 - Init START +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO comm 0x972be60 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc666fb977eaae014 - Init START +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO comm 0x9fcfea0 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc666fb977eaae014 - Init START +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO comm 0x875fec0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc666fb977eaae014 - Init START +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO comm 0x9dcafe0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc666fb977eaae014 - Init START +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO comm 0x8d79300 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc666fb977eaae014 - Init START +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO comm 0x9109680 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc666fb977eaae014 - Init START +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO comm 0x9f997a0 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc666fb977eaae014 - Init START +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO comm 0x9cf1dc0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc666fb977eaae014 - Init START +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO comm 0x872f960 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc666fb977eaae014 - Init START +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO comm 0x9b46880 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc666fb977eaae014 - Init START +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO comm 0x87c9e40 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc666fb977eaae014 - Init START +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO comm 0x9682c30 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc666fb977eaae014 - Init START +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO comm 0x89cdc40 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc666fb977eaae014 - Init START +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO comm 0xa3bcc80 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc666fb977eaae014 - Init START +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO comm 0x9280db0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc666fb977eaae014 - Init START +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO comm 0xa32f3e0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc666fb977eaae014 - Init START +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO comm 0x8dcedc0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc666fb977eaae014 - Init START +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO comm 0x8b9af40 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc666fb977eaae014 - Init START +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO comm 0x8fd4d00 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc666fb977eaae014 - Init START +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO comm 0x8c72f10 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc666fb977eaae014 - Init START +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO comm 0x96eea80 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc666fb977eaae014 - Init START +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO comm 0x9bafd50 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc666fb977eaae014 - Init START +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO comm 0x98ad880 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc666fb977eaae014 - Init START +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO comm 0x9de5d40 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc666fb977eaae014 - Init START +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO comm 0x8f1c8c0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc666fb977eaae014 - Init START +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO comm 0x98b1120 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc666fb977eaae014 - Init START +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 15/-1/-1->31->-1 [3] 15/-1/-1->31->-1 +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Trees [0] 16/-1/-1->0->-1 [1] 16/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Trees [0] 8/24/-1->16->0 [1] 8/24/-1->16->0 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Connected all rings +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Connected all rings +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Connected all rings +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Connected all rings +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Connected all rings +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Connected all rings +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Connected all rings +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Connected all rings +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Connected all rings +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Connected all rings +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Connected all rings +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Connected all rings +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Connected all rings +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Connected all rings +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Connected all rings +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Connected all rings +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Connected all rings +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Connected all rings +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Connected all rings +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070844:1071175 [7] NCCL INFO comm 0x972be60 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO Connected all trees +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656306:2656666 [0] NCCL INFO comm 0x9087740 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO Connected all trees +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO Connected all trees +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO Connected all trees +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO Connected all trees +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110736:2111063 [6] NCCL INFO comm 0x875fec0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-164-187:2110731:2111067 [1] NCCL INFO comm 0x8f1c8c0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-163-43:2656307:2656673 [1] NCCL INFO comm 0x8c78740 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO Connected all trees +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656308:2656672 [2] NCCL INFO comm 0xa137c00 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-167-9:1070839:1071170 [2] NCCL INFO comm 0x8fd4d00 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-167-9:1070843:1071174 [6] NCCL INFO comm 0x9682c30 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-164-187:2110732:2111068 [2] NCCL INFO comm 0x9de5d40 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-167-9:1070838:1071171 [1] NCCL INFO comm 0x8c72f10 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO Connected all trees +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO Connected all trees +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO Connected all trees +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO Connected all trees +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO Connected all trees +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074707:1075043 [2] NCCL INFO comm 0x9109680 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-163-43:2656312:2656668 [6] NCCL INFO comm 0x8d79300 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-163-43:2656313:2656670 [7] NCCL INFO comm 0x9dcafe0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO Connected all trees +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070840:1071173 [3] NCCL INFO comm 0x8dcedc0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO Connected all trees +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656309:2656667 [3] NCCL INFO comm 0x872f830 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-166-244:1074711:1075039 [6] NCCL INFO comm 0xa32f3e0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-164-187:2110737:2111062 [7] NCCL INFO comm 0x9fcfea0 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO Connected all trees +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656311:2656669 [5] NCCL INFO comm 0x9702210 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO Connected all trees +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110733:2111066 [3] NCCL INFO comm 0x98ad880 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-163-43:2656310:2656671 [4] NCCL INFO comm 0x9329580 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110735:2111064 [5] NCCL INFO comm 0xa3bcc80 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-167-9:1070837:1071169 [0] NCCL INFO comm 0x96eea80 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO Connected all trees +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110734:2111065 [4] NCCL INFO comm 0x9bafd50 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-167-9:1070842:1071176 [5] NCCL INFO comm 0x89cdc40 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-167-9:1070841:1071172 [4] NCCL INFO comm 0x8b9af40 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-164-187:2110730:2111069 [0] NCCL INFO comm 0x98b1120 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074708:1075040 [3] NCCL INFO comm 0x9f997a0 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074712:1075038 [7] NCCL INFO comm 0x9280db0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-166-244:1074710:1075044 [5] NCCL INFO comm 0x9b46880 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-166-244:1074706:1075042 [1] NCCL INFO comm 0x872f960 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-166-244:1074709:1075037 [4] NCCL INFO comm 0x87c9e40 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074705:1075041 [0] NCCL INFO comm 0x9cf1dc0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc666fb977eaae014 - Init COMPLETE +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO comm 0x8743820 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x66c56f39fabc3c29 - Init START +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO comm 0xa14a8c0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x66c56f39fabc3c29 - Init START +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO comm 0x8c8c700 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x66c56f39fabc3c29 - Init START +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO comm 0x909a400 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x66c56f39fabc3c29 - Init START +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO comm 0x9dddca0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x66c56f39fabc3c29 - Init START +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO comm 0x933c240 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x66c56f39fabc3c29 - Init START +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO comm 0x8d8bfc0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x66c56f39fabc3c29 - Init START +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO comm 0x9714ed0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x66c56f39fabc3c29 - Init START +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO comm 0x9fe2b60 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa1927eccc6f26a3b - Init START +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO comm 0x8772b80 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa1927eccc6f26a3b - Init START +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO comm 0x98c3de0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa1927eccc6f26a3b - Init START +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO comm 0xa3cf940 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa1927eccc6f26a3b - Init START +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO comm 0x9bc2a10 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa1927eccc6f26a3b - Init START +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO comm 0x98c0540 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa1927eccc6f26a3b - Init START +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO comm 0x9df8a00 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa1927eccc6f26a3b - Init START +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO comm 0x8f2f580 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa1927eccc6f26a3b - Init START +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO comm 0x9fac460 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5208f9921eb5621b - Init START +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO comm 0x87dcb00 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5208f9921eb5621b - Init START +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO comm 0x911c340 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5208f9921eb5621b - Init START +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO comm 0x8742620 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5208f9921eb5621b - Init START +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO comm 0x9b59540 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5208f9921eb5621b - Init START +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO comm 0x9d04a80 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5208f9921eb5621b - Init START +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO comm 0x9293a70 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5208f9921eb5621b - Init START +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO comm 0xa3420a0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5208f9921eb5621b - Init START +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO comm 0x9696920 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd6ab89002a2e3cc4 - Init START +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO comm 0x9701740 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd6ab89002a2e3cc4 - Init START +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO comm 0x973eb20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd6ab89002a2e3cc4 - Init START +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO comm 0x8c85bd0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd6ab89002a2e3cc4 - Init START +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO comm 0x89e0900 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd6ab89002a2e3cc4 - Init START +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO comm 0x8badc00 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd6ab89002a2e3cc4 - Init START +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO comm 0x8de1a80 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd6ab89002a2e3cc4 - Init START +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO comm 0x8fe79c0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd6ab89002a2e3cc4 - Init START +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Connected all rings +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Connected all rings +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Connected all rings +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Connected all rings +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Connected all rings +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Connected all rings +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Connected all rings +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Connected all rings +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Connected all rings +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070844:1071209 [7] NCCL INFO comm 0x973eb20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd6ab89002a2e3cc4 - Init COMPLETE +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070837:1071203 [0] NCCL INFO comm 0x9701740 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd6ab89002a2e3cc4 - Init COMPLETE +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070838:1071206 [1] NCCL INFO comm 0x8c85bd0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd6ab89002a2e3cc4 - Init COMPLETE +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO Connected all trees +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070839:1071204 [2] NCCL INFO comm 0x8fe79c0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd6ab89002a2e3cc4 - Init COMPLETE +ip-26-0-167-9:1070843:1071207 [6] NCCL INFO comm 0x9696920 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd6ab89002a2e3cc4 - Init COMPLETE +ip-26-0-164-187:2110730:2111097 [0] NCCL INFO comm 0x98c3de0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa1927eccc6f26a3b - Init COMPLETE +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO Connected all trees +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070840:1071208 [3] NCCL INFO comm 0x8de1a80 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd6ab89002a2e3cc4 - Init COMPLETE +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110737:2111098 [7] NCCL INFO comm 0x9fe2b60 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa1927eccc6f26a3b - Init COMPLETE +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070842:1071210 [5] NCCL INFO comm 0x89e0900 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd6ab89002a2e3cc4 - Init COMPLETE +ip-26-0-167-9:1070841:1071205 [4] NCCL INFO comm 0x8badc00 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd6ab89002a2e3cc4 - Init COMPLETE +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO Connected all trees +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Connected all rings +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Connected all rings +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Connected all rings +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Connected all rings +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Connected all rings +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110731:2111104 [1] NCCL INFO comm 0x8f2f580 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa1927eccc6f26a3b - Init COMPLETE +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Connected all rings +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Connected all rings +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Connected all rings +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO Connected all trees +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110736:2111101 [6] NCCL INFO comm 0x8772b80 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa1927eccc6f26a3b - Init COMPLETE +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO Connected all trees +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO Connected all trees +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO Connected all trees +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074705:1075072 [0] NCCL INFO comm 0x9d04a80 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5208f9921eb5621b - Init COMPLETE +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO Connected all trees +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110733:2111102 [3] NCCL INFO comm 0x98c0540 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa1927eccc6f26a3b - Init COMPLETE +ip-26-0-164-187:2110735:2111099 [5] NCCL INFO comm 0xa3cf940 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa1927eccc6f26a3b - Init COMPLETE +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110734:2111100 [4] NCCL INFO comm 0x9bc2a10 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa1927eccc6f26a3b - Init COMPLETE +ip-26-0-164-187:2110732:2111103 [2] NCCL INFO comm 0x9df8a00 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa1927eccc6f26a3b - Init COMPLETE +ip-26-0-166-244:1074712:1075077 [7] NCCL INFO comm 0x9293a70 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5208f9921eb5621b - Init COMPLETE +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074706:1075075 [1] NCCL INFO comm 0x8742620 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5208f9921eb5621b - Init COMPLETE +ip-26-0-166-244:1074711:1075074 [6] NCCL INFO comm 0xa3420a0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5208f9921eb5621b - Init COMPLETE +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO Connected all trees +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074708:1075073 [3] NCCL INFO comm 0x9fac460 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5208f9921eb5621b - Init COMPLETE +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074707:1075078 [2] NCCL INFO comm 0x911c340 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5208f9921eb5621b - Init COMPLETE +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO Connected all trees +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074710:1075079 [5] NCCL INFO comm 0x9b59540 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5208f9921eb5621b - Init COMPLETE +ip-26-0-163-43:2656306:2656700 [0] NCCL INFO comm 0x909a400 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x66c56f39fabc3c29 - Init COMPLETE +ip-26-0-166-244:1074709:1075076 [4] NCCL INFO comm 0x87dcb00 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5208f9921eb5621b - Init COMPLETE +ip-26-0-163-43:2656313:2656703 [7] NCCL INFO comm 0x9dddca0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x66c56f39fabc3c29 - Init COMPLETE +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO Connected all trees +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO Connected all trees +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656308:2656707 [2] NCCL INFO comm 0xa14a8c0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x66c56f39fabc3c29 - Init COMPLETE +ip-26-0-163-43:2656307:2656705 [1] NCCL INFO comm 0x8c8c700 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x66c56f39fabc3c29 - Init COMPLETE +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO Connected all trees +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656312:2656701 [6] NCCL INFO comm 0x8d8bfc0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x66c56f39fabc3c29 - Init COMPLETE +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO Connected all trees +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO Connected all trees +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO Connected all trees +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656309:2656704 [3] NCCL INFO comm 0x8743820 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x66c56f39fabc3c29 - Init COMPLETE +ip-26-0-163-43:2656311:2656702 [5] NCCL INFO comm 0x9714ed0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x66c56f39fabc3c29 - Init COMPLETE +ip-26-0-163-43:2656310:2656706 [4] NCCL INFO comm 0x933c240 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x66c56f39fabc3c29 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO comm 0x9f56ce0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x93bf7fc1f5ee97f9 - Init START +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO comm 0x9a44150 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbe68b0d7dae75b1b - Init START +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO comm 0xa0b4c90 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5f68ae3b534530e7 - Init START +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO comm 0x9cb58d0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x91211704300fb1e3 - Init START +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO comm 0x9aad9d0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x825436ac75a3dc08 - Init START +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO comm 0xae935e0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf7a1ffe62003bb09 - Init START +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO comm 0xb2b5090 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe33a683a707bfe4d - Init START +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO comm 0x9e83200 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x50e1c6414c9177cc - Init START +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO comm 0xb27d9f0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd53a13437b75d723 - Init START +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO comm 0xaa0f0f0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6802e6fb56477a7a - Init START +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO comm 0xa968e70 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1e11935d09ff38f1 - Init START +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO comm 0xab995a0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x687d93ed25777e6b - Init START +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO comm 0xa201390 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xff31adffe397ad2b - Init START +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO comm 0xa528280 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7c85034c7d0b48ba - Init START +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO comm 0xa061ba0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x15a4baa0c50e474f - Init START +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO comm 0xa9e6c30 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2ffd10dc5f06a6a9 - Init START +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO comm 0x9a168e0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4d01b7a752f3aa3a - Init START +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO comm 0xab93af0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7f7d3d527f8d094d - Init START +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO comm 0xa60edb0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x89416a4dae03ab11 - Init START +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO comm 0xafd6e90 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7cbcb21f1b991516 - Init START +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO comm 0xb41dae0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7ef5bcb099d445bc - Init START +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO comm 0x9a17e80 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xcff5b4a1e4cd8827 - Init START +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO comm 0xb0c89e0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9c4edd463d9c6a08 - Init START +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO comm 0xa9d6af0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8c486557ea7aa7f2 - Init START +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO comm 0xa2bc080 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2694767268b49677 - Init START +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO comm 0xae2c260 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbd7eae0f120c0376 - Init START +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO comm 0xb6a1ce0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xed9be84f8b205434 - Init START +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO comm 0xa5661a0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9e145aa54e3a65ed - Init START +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO comm 0xa128830 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9ed24ddaf1b65a03 - Init START +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO comm 0xb0b08e0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc173006364b143c8 - Init START +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO comm 0xb614b00 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xce055246947aceb5 - Init START +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO comm 0xa3ede90 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x280d2b3bed9e39c1 - Init START +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Connected all rings +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO Connected all trees +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Connected all rings +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO Connected all trees +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Connected all rings +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-9:1070842:1071250 [5] NCCL INFO comm 0x9cb58d0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x91211704300fb1e3 - Init COMPLETE +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Connected all rings +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO Connected all trees +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-166-244:1074710:1075120 [5] NCCL INFO comm 0xae2c260 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbd7eae0f120c0376 - Init COMPLETE +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070843:1071244 [6] NCCL INFO comm 0xa968e70 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1e11935d09ff38f1 - Init COMPLETE +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110731:2111142 [1] NCCL INFO comm 0xa201390 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xff31adffe397ad2b - Init COMPLETE +ip-26-0-167-9:1070838:1071248 [1] NCCL INFO comm 0x9f56ce0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x93bf7fc1f5ee97f9 - Init COMPLETE +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656309:2656750 [3] NCCL INFO comm 0x9a17e80 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xcff5b4a1e4cd8827 - Init COMPLETE +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Connected all rings +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO Connected all trees +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Connected all rings +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO Connected all trees +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-9:1070844:1071254 [7] NCCL INFO comm 0xaa0f0f0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6802e6fb56477a7a - Init COMPLETE +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO Connected all trees +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Connected all rings +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO Connected all trees +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656312:2656740 [6] NCCL INFO comm 0xa061ba0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x15a4baa0c50e474f - Init COMPLETE +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-9:1070840:1071251 [3] NCCL INFO comm 0xa0b4c90 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5f68ae3b534530e7 - Init COMPLETE +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Connected all rings +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO Connected all trees +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-9:1070837:1071247 [0] NCCL INFO comm 0xa9d6af0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8c486557ea7aa7f2 - Init COMPLETE +ip-26-0-166-244:1074706:1075116 [1] NCCL INFO comm 0x9a168e0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4d01b7a752f3aa3a - Init COMPLETE +ip-26-0-167-9:1070841:1071252 [4] NCCL INFO comm 0x9e83200 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x50e1c6414c9177cc - Init COMPLETE +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Connected all rings +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO Connected all trees +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-9:1070839:1071255 [2] NCCL INFO comm 0xa2bc080 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2694767268b49677 - Init COMPLETE +ip-26-0-164-187:2110735:2111144 [5] NCCL INFO comm 0xb6a1ce0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xed9be84f8b205434 - Init COMPLETE +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Connected all rings +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO Connected all trees +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-43:2656306:2656743 [0] NCCL INFO comm 0xa528280 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7c85034c7d0b48ba - Init COMPLETE +ip-26-0-163-43:2656311:2656748 [5] NCCL INFO comm 0xa9e6c30 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2ffd10dc5f06a6a9 - Init COMPLETE +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Connected all rings +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO Connected all trees +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Connected all rings +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO Connected all trees +ip-26-0-164-187:2110732:2111149 [2] NCCL INFO comm 0xb0c89e0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9c4edd463d9c6a08 - Init COMPLETE +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Connected all rings +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO Connected all trees +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-166-244:1074711:1075123 [6] NCCL INFO comm 0xb614b00 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xce055246947aceb5 - Init COMPLETE +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-166-244:1074708:1075112 [3] NCCL INFO comm 0xb27d9f0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd53a13437b75d723 - Init COMPLETE +ip-26-0-166-244:1074707:1075124 [2] NCCL INFO comm 0xa3ede90 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x280d2b3bed9e39c1 - Init COMPLETE +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074709:1075113 [4] NCCL INFO comm 0x9aad9d0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x825436ac75a3dc08 - Init COMPLETE +ip-26-0-164-187:2110733:2111148 [3] NCCL INFO comm 0xab93af0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7f7d3d527f8d094d - Init COMPLETE +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110730:2111146 [0] NCCL INFO comm 0xab995a0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x687d93ed25777e6b - Init COMPLETE +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Connected all rings +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO Connected all trees +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-164-187:2110737:2111137 [7] NCCL INFO comm 0xb2b5090 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe33a683a707bfe4d - Init COMPLETE +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Connected all rings +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO Connected all trees +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-166-244:1074712:1075119 [7] NCCL INFO comm 0xa5661a0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9e145aa54e3a65ed - Init COMPLETE +ip-26-0-166-244:1074705:1075118 [0] NCCL INFO comm 0xafd6e90 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7cbcb21f1b991516 - Init COMPLETE +ip-26-0-164-187:2110736:2111136 [6] NCCL INFO comm 0x9a44150 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbe68b0d7dae75b1b - Init COMPLETE +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110734:2111145 [4] NCCL INFO comm 0xae935e0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf7a1ffe62003bb09 - Init COMPLETE +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO comm 0xb453c90 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO comm 0xbcf9cf0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO comm 0xc0e6970 rank 6 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO comm 0xafaadd0 rank 11 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO comm 0xb870f60 rank 10 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO comm 0xb5d84d0 rank 5 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO comm 0xa6fa5e0 rank 14 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO comm 0xaaf9840 rank 13 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO comm 0xac46310 rank 4 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO comm 0xa99bcb0 rank 12 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO comm 0xa45b520 rank 8 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO comm 0xbcc2a50 rank 9 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO comm 0xb3ad920 rank 15 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO comm 0xa488b50 rank 7 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO comm 0xb8d81a0 rank 6 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO comm 0xc0598f0 rank 11 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO comm 0xa4f2860 rank 10 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO comm 0xa8c8220 rank 14 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO comm 0xad00ee0 rank 13 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO comm 0xb5de0a0 rank 4 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO comm 0xbb0d0a0 rank 5 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO comm 0xae32720 rank 9 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO comm 0xba1c040 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO comm 0xb41bba0 rank 12 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Connected all rings +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO Connected all trees +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Connected all rings +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO Connected all trees +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-43:2656310:2656744 [4] NCCL INFO comm 0xa60edb0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x89416a4dae03ab11 - Init COMPLETE +ip-26-0-163-43:2656307:2656751 [1] NCCL INFO comm 0xa128830 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9ed24ddaf1b65a03 - Init COMPLETE +ip-26-0-163-43:2656313:2656752 [7] NCCL INFO comm 0xb0b08e0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc173006364b143c8 - Init COMPLETE +ip-26-0-163-43:2656308:2656749 [2] NCCL INFO comm 0xb41dae0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7ef5bcb099d445bc - Init COMPLETE +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO comm 0xab6d420 rank 0 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO comm 0xbaf5740 rank 3 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO comm 0xa45cba0 rank 1 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO comm 0xb42ba50 rank 2 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xab5b6082ecdc6e3a - Init START +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO comm 0xaf6d100 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO comm 0xbe62600 rank 1 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO comm 0xb053520 rank 2 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO comm 0xaaa6af0 rank 3 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc5d2d86eb8477ad1 - Init START +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Trees [0] 8/-1/-1->0->-1 [1] 8/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Trees [0] 4/12/-1->8->0 [1] 4/12/-1->8->0 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 7/-1/-1->15->-1 [3] 7/-1/-1->15->-1 +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 7/-1/-1->15->-1 [3] 7/-1/-1->15->-1 +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Trees [0] 8/-1/-1->0->-1 [1] 8/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Trees [0] 4/12/-1->8->0 [1] 4/12/-1->8->0 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 00/0 : 14[5] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 00/0 : 13[3] -> 14[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 01/0 : 14[5] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 01/0 : 13[3] -> 14[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 02/0 : 14[5] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 02/0 : 13[3] -> 14[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 03/0 : 14[5] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 03/0 : 13[3] -> 14[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 00/0 : 14[5] -> 15[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 00/0 : 12[1] -> 13[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 01/0 : 15[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 00/0 : 11[7] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 01/0 : 14[5] -> 15[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 02/0 : 15[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 01/0 : 12[1] -> 13[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 02/0 : 14[5] -> 15[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 03/0 : 15[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 01/0 : 11[7] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 02/0 : 12[1] -> 13[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 03/0 : 14[5] -> 15[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 02/0 : 11[7] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 03/0 : 12[1] -> 13[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 00/0 : 13[3] -> 14[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 03/0 : 11[7] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 01/0 : 13[3] -> 14[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 00/0 : 12[1] -> 13[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 02/0 : 13[3] -> 14[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 01/0 : 12[1] -> 13[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 00/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 00/0 : 8[1] -> 9[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 03/0 : 13[3] -> 14[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 02/0 : 12[1] -> 13[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 01/0 : 8[1] -> 9[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 01/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 03/0 : 12[1] -> 13[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 02/0 : 8[1] -> 9[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 02/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 03/0 : 8[1] -> 9[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 03/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 00/0 : 9[3] -> 10[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 01/0 : 9[3] -> 10[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 02/0 : 9[3] -> 10[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 03/0 : 9[3] -> 10[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 00/0 : 10[5] -> 11[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 00/0 : 7[7] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 01/0 : 10[5] -> 11[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 01/0 : 7[7] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 02/0 : 10[5] -> 11[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 03/0 : 10[5] -> 11[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 02/0 : 7[7] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 00/0 : 11[7] -> 12[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 03/0 : 7[7] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 00/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 01/0 : 11[7] -> 12[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 00/0 : 8[1] -> 9[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 02/0 : 11[7] -> 12[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 01/0 : 8[1] -> 9[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 03/0 : 11[7] -> 12[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 01/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 02/0 : 8[1] -> 9[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 02/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 03/0 : 8[1] -> 9[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 03/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 00/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 01/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 00/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 02/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 01/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 03/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 02/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 00/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 00/0 : 9[3] -> 10[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 03/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 01/0 : 9[3] -> 10[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 01/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 00/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 02/0 : 9[3] -> 10[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 02/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 01/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 03/0 : 9[3] -> 10[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 03/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 02/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 00/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 00/0 : 10[5] -> 11[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 00/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 03/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 00/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 01/0 : 10[5] -> 11[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 01/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 01/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 02/0 : 10[5] -> 11[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 01/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 02/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 02/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 02/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 03/0 : 10[5] -> 11[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 03/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 03/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 03/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 00/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 01/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 02/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 03/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 00/0 : 15[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 01/0 : 15[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 02/0 : 15[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 03/0 : 15[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 00/0 : 12[0] -> 13[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 01/0 : 12[0] -> 13[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 02/0 : 12[0] -> 13[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 00/0 : 14[4] -> 15[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 03/0 : 12[0] -> 13[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 00/0 : 13[2] -> 14[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 00/0 : 13[2] -> 14[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 01/0 : 14[4] -> 15[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 01/0 : 13[2] -> 14[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 02/0 : 14[4] -> 15[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 02/0 : 13[2] -> 14[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 01/0 : 13[2] -> 14[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 03/0 : 13[2] -> 14[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 03/0 : 14[4] -> 15[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 02/0 : 13[2] -> 14[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 00/0 : 15[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 03/0 : 13[2] -> 14[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 00/0 : 11[6] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 01/0 : 15[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 00/0 : 14[4] -> 15[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 02/0 : 15[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 01/0 : 14[4] -> 15[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 03/0 : 15[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 02/0 : 14[4] -> 15[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 01/0 : 11[6] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 03/0 : 14[4] -> 15[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 02/0 : 11[6] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 03/0 : 11[6] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 00/0 : 12[0] -> 13[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 01/0 : 12[0] -> 13[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 02/0 : 12[0] -> 13[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 00/0 : 10[4] -> 11[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 03/0 : 12[0] -> 13[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 01/0 : 10[4] -> 11[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 02/0 : 10[4] -> 11[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 03/0 : 10[4] -> 11[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 00/0 : 11[6] -> 12[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 01/0 : 11[6] -> 12[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 02/0 : 11[6] -> 12[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 03/0 : 11[6] -> 12[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 00/0 : 8[0] -> 9[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 00/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 01/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 01/0 : 8[0] -> 9[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 02/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 02/0 : 8[0] -> 9[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 03/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 03/0 : 8[0] -> 9[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 00/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 00/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 00/0 : 9[2] -> 10[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 01/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 01/0 : 9[2] -> 10[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 01/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 02/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 02/0 : 9[2] -> 10[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 03/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 02/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 03/0 : 9[2] -> 10[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 03/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 00/0 : 9[2] -> 10[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 01/0 : 9[2] -> 10[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 02/0 : 9[2] -> 10[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 03/0 : 9[2] -> 10[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 00/0 : 7[6] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 00/0 : 10[4] -> 11[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 01/0 : 10[4] -> 11[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 01/0 : 7[6] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 02/0 : 10[4] -> 11[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 02/0 : 7[6] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 03/0 : 10[4] -> 11[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 03/0 : 7[6] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 00/0 : 15[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 01/0 : 15[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 02/0 : 15[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 03/0 : 15[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 00/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 01/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 02/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 03/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 00/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 00/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 00/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 01/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 01/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 01/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 02/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 02/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 02/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 00/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 03/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 03/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 03/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 00/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 01/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 00/0 : 7[6] -> 8[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 01/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 02/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 01/0 : 7[6] -> 8[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 02/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 03/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 02/0 : 7[6] -> 8[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 03/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 00/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 03/0 : 7[6] -> 8[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 01/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 02/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 03/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Connected all rings +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 00/0 : 4[1] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 01/0 : 4[1] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Connected all rings +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 00/0 : 2[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 01/0 : 2[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 00/0 : 4[1] -> 6[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 01/0 : 4[1] -> 6[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 00/0 : 6[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 01/0 : 6[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Connected all rings +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 02/0 : 3[7] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 03/0 : 3[7] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 00/0 : 12[1] -> 14[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 01/0 : 12[1] -> 14[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Connected all rings +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 00/0 : 10[5] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 01/0 : 10[5] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 00/0 : 12[1] -> 14[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 02/0 : 11[7] -> 13[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 03/0 : 11[7] -> 13[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 01/0 : 12[1] -> 14[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 00/0 : 14[5] -> 12[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 01/0 : 14[5] -> 12[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Connected all rings +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Connected all rings +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Connected all rings +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Connected all rings +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 00/0 : 4[0] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 00/0 : 2[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 01/0 : 4[0] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 00/0 : 8[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 01/0 : 2[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 01/0 : 8[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 00/0 : 0[1] -> 8[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 01/0 : 0[1] -> 8[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Connected all rings +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 02/0 : 3[6] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 03/0 : 3[6] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 00/0 : 4[1] -> 8[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 00/0 : 4[1] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 01/0 : 4[1] -> 8[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 01/0 : 4[1] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Connected all rings +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 02/0 : 1[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 03/0 : 1[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 02/0 : 3[6] -> 5[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 03/0 : 3[6] -> 5[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 02/0 : 5[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 03/0 : 5[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Connected all rings +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 00/0 : 2[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 01/0 : 2[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 00/0 : 4[0] -> 6[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Connected all rings +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 01/0 : 4[0] -> 6[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 00/0 : 12[0] -> 14[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 01/0 : 12[0] -> 14[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 00/0 : 6[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 01/0 : 6[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 02/0 : 11[6] -> 13[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 03/0 : 11[6] -> 13[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Connected all rings +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 00/0 : 10[5] -> 12[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 01/0 : 10[5] -> 12[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 00/0 : 12[1] -> 10[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 00/0 : 8[1] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 01/0 : 12[1] -> 10[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 01/0 : 8[1] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Connected all rings +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 00/0 : 2[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Connected all rings +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 01/0 : 2[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 02/0 : 1[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 03/0 : 1[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Connected all rings +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 02/0 : 3[7] -> 5[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 00/0 : 10[4] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 03/0 : 3[7] -> 5[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 01/0 : 10[4] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 00/0 : 12[0] -> 14[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 00/0 : 4[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 00/0 : 4[0] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 01/0 : 12[0] -> 14[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 01/0 : 4[0] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Connected all rings +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 02/0 : 3[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 03/0 : 3[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 02/0 : 7[6] -> 15[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 03/0 : 7[6] -> 15[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 00/0 : 14[4] -> 12[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 02/0 : 15[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 01/0 : 14[4] -> 12[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 03/0 : 15[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 01/0 : 4[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 02/0 : 5[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 03/0 : 5[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Connected all rings +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 02/0 : 3[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 03/0 : 3[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Connected all rings +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 02/0 : 9[3] -> 11[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 03/0 : 9[3] -> 11[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Connected all rings +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 02/0 : 7[6] -> 11[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 03/0 : 7[6] -> 11[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 02/0 : 9[2] -> 11[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 03/0 : 9[2] -> 11[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 02/0 : 11[6] -> 13[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 03/0 : 11[6] -> 13[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 02/0 : 13[2] -> 11[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 03/0 : 13[2] -> 11[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 02/0 : 9[3] -> 11[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 03/0 : 9[3] -> 11[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 02/0 : 11[7] -> 13[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 03/0 : 11[7] -> 13[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 02/0 : 11[7] -> 9[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 03/0 : 11[7] -> 9[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 02/0 : 13[3] -> 11[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 03/0 : 13[3] -> 11[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 02/0 : 11[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 03/0 : 11[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 02/0 : 11[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 03/0 : 11[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 02/0 : 13[3] -> 11[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 03/0 : 13[3] -> 11[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 02/0 : 11[7] -> 9[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 03/0 : 11[7] -> 9[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 02/0 : 5[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 03/0 : 5[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 00/0 : 10[5] -> 9[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 00/0 : 11[7] -> 10[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 01/0 : 10[5] -> 9[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Channel 01/0 : 11[7] -> 10[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 02/0 : 10[5] -> 9[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 00/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 03/0 : 10[5] -> 9[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 02/0 : 9[3] -> 8[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 01/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 02/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Channel 03/0 : 9[3] -> 8[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 03/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 00/0 : 4[1] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 02/0 : 5[3] -> 4[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Channel 03/0 : 5[3] -> 4[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 01/0 : 4[1] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 00/0 : 8[1] -> 12[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 01/0 : 8[1] -> 12[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 00/0 : 8[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 01/0 : 8[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 00/0 : 0[1] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 00/0 : 12[1] -> 8[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 01/0 : 0[1] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 01/0 : 12[1] -> 8[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 00/0 : 8[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 01/0 : 8[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 00/0 : 12[1] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 01/0 : 12[1] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 00/0 : 8[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 01/0 : 8[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 00/0 : 14[5] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 01/0 : 14[5] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 02/0 : 9[3] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 00/0 : 6[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 00/0 : 12[1] -> 10[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Channel 03/0 : 9[3] -> 8[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 01/0 : 6[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 01/0 : 12[1] -> 10[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 00/0 : 4[1] -> 2[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 01/0 : 4[1] -> 2[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 00/0 : 7[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 02/0 : 13[3] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 01/0 : 7[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Channel 03/0 : 13[3] -> 12[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 00/0 : 11[7] -> 10[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 01/0 : 11[7] -> 10[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 00/0 : 3[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 00/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 02/0 : 5[3] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 00/0 : 10[5] -> 9[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 01/0 : 3[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 01/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Channel 03/0 : 5[3] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 01/0 : 10[5] -> 9[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 02/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 02/0 : 10[5] -> 9[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Channel 03/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Channel 03/0 : 10[5] -> 9[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 00/0 : 14[5] -> 13[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 01/0 : 14[5] -> 13[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 02/0 : 14[5] -> 13[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 03/0 : 14[5] -> 13[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 02/0 : 13[3] -> 12[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Channel 03/0 : 13[3] -> 12[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 00/0 : 15[7] -> 14[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 00/0 : 10[4] -> 12[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 01/0 : 10[4] -> 12[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 00/0 : 12[0] -> 10[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 01/0 : 12[0] -> 10[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 00/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 01/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 02/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 03/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 01/0 : 15[7] -> 14[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 00/0 : 14[5] -> 13[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 01/0 : 14[5] -> 13[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 02/0 : 14[5] -> 13[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Channel 03/0 : 14[5] -> 13[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 02/0 : 9[2] -> 11[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 03/0 : 9[2] -> 11[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 02/0 : 7[6] -> 11[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 02/0 : 11[6] -> 9[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 03/0 : 7[6] -> 11[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 03/0 : 11[6] -> 9[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 02/0 : 11[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 03/0 : 11[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 02/0 : 15[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 02/0 : 13[2] -> 11[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 03/0 : 13[2] -> 11[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 02/0 : 11[6] -> 9[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 03/0 : 11[6] -> 9[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 03/0 : 15[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 02/0 : 7[6] -> 15[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 03/0 : 7[6] -> 15[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 02/0 : 11[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 03/0 : 11[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 00/0 : 7[6] -> 6[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Channel 01/0 : 7[6] -> 6[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 00/0 : 10[4] -> 9[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 00/0 : 11[6] -> 10[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Channel 01/0 : 11[6] -> 10[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 01/0 : 10[4] -> 9[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 00/0 : 15[6] -> 14[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Channel 01/0 : 15[6] -> 14[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 00/0 : 14[4] -> 13[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 01/0 : 14[4] -> 13[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 02/0 : 14[4] -> 13[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 02/0 : 10[4] -> 9[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 03/0 : 10[4] -> 9[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 02/0 : 9[2] -> 8[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Channel 03/0 : 9[2] -> 8[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 03/0 : 14[4] -> 13[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 02/0 : 13[2] -> 12[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Channel 03/0 : 13[2] -> 12[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 02/0 : 5[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 03/0 : 5[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 00/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 01/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 02/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 03/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 02/0 : 5[2] -> 4[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Channel 03/0 : 5[2] -> 4[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 00/0 : 4[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 01/0 : 4[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 00/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 01/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 02/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 03/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 00/0 : 12[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 01/0 : 12[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 00/0 : 12[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 01/0 : 12[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 00/0 : 14[4] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 01/0 : 14[4] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 02/0 : 9[2] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 00/0 : 12[0] -> 10[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Channel 03/0 : 9[2] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 01/0 : 12[0] -> 10[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 00/0 : 15[6] -> 14[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 01/0 : 15[6] -> 14[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 02/0 : 13[2] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 00/0 : 11[6] -> 10[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 00/0 : 14[4] -> 13[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 01/0 : 11[6] -> 10[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Channel 03/0 : 13[2] -> 12[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 00/0 : 10[4] -> 9[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 01/0 : 14[4] -> 13[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 02/0 : 14[4] -> 13[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 01/0 : 10[4] -> 9[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Channel 03/0 : 14[4] -> 13[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 02/0 : 10[4] -> 9[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Channel 03/0 : 10[4] -> 9[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 00/0 : 6[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 01/0 : 6[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 00/0 : 4[0] -> 2[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 01/0 : 4[0] -> 2[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 00/0 : 7[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 01/0 : 7[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 02/0 : 5[2] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 00/0 : 3[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 00/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Channel 03/0 : 5[2] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 01/0 : 3[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 01/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 02/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Channel 03/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO Connected all trees +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656306:2656777 [0] NCCL INFO comm 0xaf6d100 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-167-9:1070844:1071274 [7] NCCL INFO comm 0xb453c90 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO Connected all trees +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656307:2656772 [1] NCCL INFO comm 0xab6d420 rank 0 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-167-9:1070843:1071278 [6] NCCL INFO comm 0xb3ad920 rank 15 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO Connected all trees +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO Connected all trees +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO Connected all trees +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074708:1075144 [3] NCCL INFO comm 0xbcc2a50 rank 9 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO Connected all trees +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110733:2111168 [3] NCCL INFO comm 0xb5d84d0 rank 5 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-166-244:1074709:1075146 [4] NCCL INFO comm 0xa4f2860 rank 10 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO Connected all trees +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656309:2656773 [3] NCCL INFO comm 0xa45cba0 rank 1 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-164-187:2110734:2111173 [4] NCCL INFO comm 0xb8d81a0 rank 6 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-163-43:2656311:2656774 [5] NCCL INFO comm 0xb42ba50 rank 2 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO Connected all trees +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656310:2656780 [4] NCCL INFO comm 0xb053520 rank 2 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-163-43:2656308:2656778 [2] NCCL INFO comm 0xbe62600 rank 1 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO Connected all trees +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656313:2656775 [7] NCCL INFO comm 0xbaf5740 rank 3 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-167-9:1070842:1071275 [5] NCCL INFO comm 0xa6fa5e0 rank 14 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO Connected all trees +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO Connected all trees +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074710:1075143 [5] NCCL INFO comm 0xb870f60 rank 10 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO Connected all trees +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO Connected all trees +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656312:2656779 [6] NCCL INFO comm 0xaaa6af0 rank 3 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-166-244:1074712:1075142 [7] NCCL INFO comm 0xafaadd0 rank 11 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO Connected all trees +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110737:2111170 [7] NCCL INFO comm 0xbcf9cf0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-164-187:2110735:2111169 [5] NCCL INFO comm 0xc0e6970 rank 6 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110732:2111172 [2] NCCL INFO comm 0xbb0d0a0 rank 5 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-164-187:2110731:2111167 [1] NCCL INFO comm 0xac46310 rank 4 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-167-9:1070841:1071279 [4] NCCL INFO comm 0xa8c8220 rank 14 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-166-244:1074706:1075145 [1] NCCL INFO comm 0xa45b520 rank 8 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO Connected all trees +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074711:1075147 [6] NCCL INFO comm 0xc0598f0 rank 11 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070840:1071276 [3] NCCL INFO comm 0xaaf9840 rank 13 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-167-9:1070838:1071277 [1] NCCL INFO comm 0xa99bcb0 rank 12 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xab5b6082ecdc6e3a - Init COMPLETE +ip-26-0-164-187:2110736:2111174 [6] NCCL INFO comm 0xa488b50 rank 7 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO Connected all trees +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074707:1075148 [2] NCCL INFO comm 0xae32720 rank 9 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074705:1075149 [0] NCCL INFO comm 0xba1c040 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-164-187:2110730:2111171 [0] NCCL INFO comm 0xb5de0a0 rank 4 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-167-9:1070839:1071281 [2] NCCL INFO comm 0xad00ee0 rank 13 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-167-9:1070837:1071280 [0] NCCL INFO comm 0xb41bba0 rank 12 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc5d2d86eb8477ad1 - Init COMPLETE +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO comm 0xaf8b360 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x62ee72480d87e3d3 - Init START +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO comm 0xab87d50 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x62ee72480d87e3d3 - Init START +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO comm 0xbab0b90 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4a43799e051bfb7 - Init START +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO comm 0xc2bea00 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4a43799e051bfb7 - Init START +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO comm 0xae1de10 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb2d0e2947588dde8 - Init START +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO comm 0xb7b6620 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb2d0e2947588dde8 - Init START +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO comm 0xa633680 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5e792e71ffeefaea - Init START +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO comm 0xbbf48f0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5e792e71ffeefaea - Init START +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO comm 0xab74da0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6c2a1f2d71c9cb2f - Init START +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO comm 0xb5f43d0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6c2a1f2d71c9cb2f - Init START +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO comm 0xa634890 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2dabe7bb14da34d0 - Init START +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO comm 0xc039a00 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2dabe7bb14da34d0 - Init START +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO comm 0xa8d5ba0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa8f6f4b3707682cb - Init START +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO comm 0xa8e3640 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa8f6f4b3707682cb - Init START +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO comm 0xb586c70 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb4f8b93ceffff548 - Init START +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO comm 0xb46f2f0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb4f8b93ceffff548 - Init START +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO comm 0xbce5f80 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x89a5c2bae28749cd - Init START +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO comm 0xb7b04c0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x89a5c2bae28749cd - Init START +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Using network Libfabric +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO comm 0xbe9a6f0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1000fceb5598d53 - Init START +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO comm 0xb00ace0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1000fceb5598d53 - Init START +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Using network Libfabric +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO comm 0xbccd9c0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xee95302585c2933d - Init START +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO comm 0xac7edb0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xee95302585c2933d - Init START +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO comm 0xa6611b0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x13531981a3d6dc65 - Init START +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO comm 0xbed26e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x13531981a3d6dc65 - Init START +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO comm 0xb22ba40 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x58627f310cc260f0 - Init START +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO comm 0xb602ed0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x58627f310cc260f0 - Init START +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO comm 0xb183430 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x15600ed997f5f6ad - Init START +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO comm 0xc231be0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x15600ed997f5f6ad - Init START +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO comm 0xba49a90 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd4e6023983e573d7 - Init START +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO comm 0xa6c9fc0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd4e6023983e573d7 - Init START +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO comm 0xacd0c40 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x507b98d1eb1a99c1 - Init START +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO comm 0xaed96d0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x507b98d1eb1a99c1 - Init START +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Connected all rings +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO Connected all trees +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Connected all rings +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO Connected all trees +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070838:1071305 [1] NCCL INFO comm 0xab74da0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6c2a1f2d71c9cb2f - Init COMPLETE +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Connected all rings +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO Connected all trees +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Connected all rings +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO Connected all trees +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Connected all rings +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO Connected all trees +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Connected all rings +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO Connected all trees +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070837:1071304 [0] NCCL INFO comm 0xb5f43d0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6c2a1f2d71c9cb2f - Init COMPLETE +ip-26-0-164-187:2110737:2111205 [7] NCCL INFO comm 0xbed26e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x13531981a3d6dc65 - Init COMPLETE +ip-26-0-164-187:2110736:2111203 [6] NCCL INFO comm 0xa6611b0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x13531981a3d6dc65 - Init COMPLETE +ip-26-0-164-187:2110731:2111201 [1] NCCL INFO comm 0xae1de10 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb2d0e2947588dde8 - Init COMPLETE +ip-26-0-164-187:2110733:2111207 [3] NCCL INFO comm 0xb7b04c0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x89a5c2bae28749cd - Init COMPLETE +ip-26-0-164-187:2110730:2111199 [0] NCCL INFO comm 0xb7b6620 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb2d0e2947588dde8 - Init COMPLETE +ip-26-0-164-187:2110732:2111206 [2] NCCL INFO comm 0xbce5f80 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x89a5c2bae28749cd - Init COMPLETE +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Connected all rings +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO Connected all trees +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Connected all rings +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO Connected all trees +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110734:2111198 [4] NCCL INFO comm 0xbab0b90 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4a43799e051bfb7 - Init COMPLETE +ip-26-0-167-9:1070841:1071309 [4] NCCL INFO comm 0xa8e3640 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa8f6f4b3707682cb - Init COMPLETE +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-187:2110735:2111200 [5] NCCL INFO comm 0xc2bea00 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4a43799e051bfb7 - Init COMPLETE +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Connected all rings +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1070842:1071311 [5] NCCL INFO comm 0xa8d5ba0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa8f6f4b3707682cb - Init COMPLETE +ip-26-0-167-9:1070839:1071313 [2] NCCL INFO comm 0xaed96d0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x507b98d1eb1a99c1 - Init COMPLETE +ip-26-0-167-9:1070840:1071314 [3] NCCL INFO comm 0xacd0c40 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x507b98d1eb1a99c1 - Init COMPLETE +ip-26-0-167-9:1070844:1071310 [7] NCCL INFO comm 0xb46f2f0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb4f8b93ceffff548 - Init COMPLETE +ip-26-0-167-9:1070843:1071308 [6] NCCL INFO comm 0xb586c70 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb4f8b93ceffff548 - Init COMPLETE +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074712:1075181 [7] NCCL INFO comm 0xb183430 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x15600ed997f5f6ad - Init COMPLETE +ip-26-0-166-244:1074711:1075178 [6] NCCL INFO comm 0xc231be0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x15600ed997f5f6ad - Init COMPLETE +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1074710:1075182 [5] NCCL INFO comm 0xba49a90 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd4e6023983e573d7 - Init COMPLETE +ip-26-0-166-244:1074708:1075179 [3] NCCL INFO comm 0xbe9a6f0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1000fceb5598d53 - Init COMPLETE +ip-26-0-166-244:1074706:1075173 [1] NCCL INFO comm 0xa633680 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5e792e71ffeefaea - Init COMPLETE +ip-26-0-166-244:1074707:1075176 [2] NCCL INFO comm 0xb00ace0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1000fceb5598d53 - Init COMPLETE +ip-26-0-166-244:1074709:1075180 [4] NCCL INFO comm 0xa6c9fc0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd4e6023983e573d7 - Init COMPLETE +ip-26-0-166-244:1074705:1075172 [0] NCCL INFO comm 0xbbf48f0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5e792e71ffeefaea - Init COMPLETE +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Connected all rings +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO Connected all trees +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Connected all rings +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO Connected all trees +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656306:2656803 [0] NCCL INFO comm 0xaf8b360 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x62ee72480d87e3d3 - Init COMPLETE +ip-26-0-163-43:2656307:2656804 [1] NCCL INFO comm 0xab87d50 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x62ee72480d87e3d3 - Init COMPLETE +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Connected all rings +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO Connected all trees +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Connected all rings +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO Connected all trees +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Connected all rings +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO Connected all trees +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Connected all rings +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO Connected all trees +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Connected all rings +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO Connected all trees +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656309:2656807 [3] NCCL INFO comm 0xa634890 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2dabe7bb14da34d0 - Init COMPLETE +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Connected all rings +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO Connected all trees +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-43:2656308:2656806 [2] NCCL INFO comm 0xc039a00 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2dabe7bb14da34d0 - Init COMPLETE +ip-26-0-163-43:2656312:2656810 [6] NCCL INFO comm 0xac7edb0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xee95302585c2933d - Init COMPLETE +ip-26-0-163-43:2656311:2656813 [5] NCCL INFO comm 0xb602ed0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x58627f310cc260f0 - Init COMPLETE +ip-26-0-163-43:2656313:2656812 [7] NCCL INFO comm 0xbccd9c0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xee95302585c2933d - Init COMPLETE +ip-26-0-163-43:2656310:2656811 [4] NCCL INFO comm 0xb22ba40 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x58627f310cc260f0 - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 1 has a total capacty of 79.33 GiB of which 1.08 GiB is free. Including non-PyTorch memory, this process has 78.24 GiB memory in use. Of the allocated memory 74.30 GiB is allocated by PyTorch, and 2.21 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self.act(gate_states) * up_states +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 0 has a total capacty of 79.33 GiB of which 1.19 GiB is free. Including non-PyTorch memory, this process has 78.12 GiB memory in use. Of the allocated memory 74.30 GiB is allocated by PyTorch, and 2.21 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +12/27/2024 17:35:39 [ERROR|DP=0|PP=14|TP=0|ip-26-0-167-9]: [RECV META] Rank 14: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=14|TP=0|ip-26-0-167-9]: [RECV META] Rank 14: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=14|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=14|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=13|TP=0|ip-26-0-167-9]: [RECV META] Rank 13: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=14|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=12|TP=0|ip-26-0-167-9]: [RECV META] Rank 12: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=14|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=15|TP=0|ip-26-0-167-9]: [RECV META] Rank 15: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=13|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=13|TP=0|ip-26-0-167-9]: [RECV META] Rank 13: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=12|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=13|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=15|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=12|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=5|TP=0|ip-26-0-164-187]: [RECV META] Rank 5: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=15|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=13|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=13|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=12|TP=0|ip-26-0-167-9]: [RECV META] Rank 12: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=5|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=5|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=6|TP=0|ip-26-0-164-187]: [RECV META] Rank 6: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=12|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=15|TP=0|ip-26-0-167-9]: [RECV META] Rank 15: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=12|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=6|TP=0|ip-26-0-164-187]: [RECV META] Rank 6: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=15|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=15|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=6|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=6|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=6|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=6|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=7|TP=0|ip-26-0-164-187]: [RECV META] Rank 7: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=4|TP=0|ip-26-0-164-187]: [RECV META] Rank 4: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=7|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=5|TP=0|ip-26-0-164-187]: [RECV META] Rank 5: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=7|TP=0|ip-26-0-164-187]: [RECV META] Rank 7: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=7|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=4|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=5|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=7|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=4|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=5|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=7|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=9|TP=0|ip-26-0-166-244]: [RECV META] Rank 9: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=9|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=9|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=10|TP=0|ip-26-0-166-244]: [RECV META] Rank 10: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=9|TP=0|ip-26-0-166-244]: [RECV META] Rank 9: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=10|TP=0|ip-26-0-166-244]: [RECV META] Rank 10: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=11|TP=0|ip-26-0-166-244]: [RECV META] Rank 11: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=8|TP=0|ip-26-0-166-244]: [RECV META] Rank 8: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=8|TP=0|ip-26-0-166-244]: [RECV META] Rank 8: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=10|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=9|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=11|TP=0|ip-26-0-166-244]: [RECV META] Rank 11: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=10|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=8|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=11|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=8|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=10|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=9|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=8|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=10|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=11|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=8|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=11|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=11|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=4|TP=0|ip-26-0-164-187]: [RECV META] Rank 4: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=4|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=4|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=2|TP=0|ip-26-0-163-43]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=3|TP=0|ip-26-0-163-43]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=2|TP=0|ip-26-0-163-43]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=2|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=1|TP=0|ip-26-0-163-43]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=2|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=3|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=2|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=3|TP=0|ip-26-0-163-43]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=3|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=2|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=1|TP=0|ip-26-0-163-43]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=1|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=1|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=3|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=1|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=3|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=1|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=14|TP=0|ip-26-0-167-9]: [RECV META] Rank 14: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=14|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=14|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=12|TP=0|ip-26-0-167-9]: [RECV META] Rank 12: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=14|TP=0|ip-26-0-167-9]: [RECV META] Rank 14: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=12|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=13|TP=0|ip-26-0-167-9]: [RECV META] Rank 13: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=15|TP=0|ip-26-0-167-9]: [RECV META] Rank 15: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=14|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=12|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=13|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=14|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=15|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=13|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=15|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=5|TP=0|ip-26-0-164-187]: [RECV META] Rank 5: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=13|TP=0|ip-26-0-167-9]: [RECV META] Rank 13: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=13|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=5|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=13|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=5|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=12|TP=0|ip-26-0-167-9]: [RECV META] Rank 12: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=12|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=12|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=6|TP=0|ip-26-0-164-187]: [RECV META] Rank 6: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=15|TP=0|ip-26-0-167-9]: [RECV META] Rank 15: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=6|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=6|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=15|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=15|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=7|TP=0|ip-26-0-164-187]: [RECV META] Rank 7: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=6|TP=0|ip-26-0-164-187]: [RECV META] Rank 6: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=5|TP=0|ip-26-0-164-187]: [RECV META] Rank 5: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=7|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=6|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=7|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=5|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=7|TP=0|ip-26-0-164-187]: [RECV META] Rank 7: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=4|TP=0|ip-26-0-164-187]: [RECV META] Rank 4: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=5|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=6|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=7|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=4|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=7|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=4|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=9|TP=0|ip-26-0-166-244]: [RECV META] Rank 9: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=9|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=9|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=10|TP=0|ip-26-0-166-244]: [RECV META] Rank 10: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=8|TP=0|ip-26-0-166-244]: [RECV META] Rank 8: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=8|TP=0|ip-26-0-166-244]: [RECV META] Rank 8: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=10|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=11|TP=0|ip-26-0-166-244]: [RECV META] Rank 11: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=9|TP=0|ip-26-0-166-244]: [RECV META] Rank 9: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=8|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=8|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=10|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=11|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=10|TP=0|ip-26-0-166-244]: [RECV META] Rank 10: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=9|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=8|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=8|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=11|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=9|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=11|TP=0|ip-26-0-166-244]: [RECV META] Rank 11: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=10|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=10|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=11|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=11|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=4|TP=0|ip-26-0-164-187]: [RECV META] Rank 4: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=4|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=4|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=2|TP=0|ip-26-0-163-43]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=2|TP=0|ip-26-0-163-43]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=2|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=2|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=2|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=2|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=1|TP=0|ip-26-0-163-43]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=3|TP=0|ip-26-0-163-43]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=1|TP=0|ip-26-0-163-43]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=3|TP=0|ip-26-0-163-43]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=1|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=3|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=1|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=1|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=3|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=3|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=1|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=3|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=14|TP=0|ip-26-0-167-9]: [RECV META] Rank 14: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=14|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=12|TP=0|ip-26-0-167-9]: [RECV META] Rank 12: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=14|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=12|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=15|TP=0|ip-26-0-167-9]: [RECV META] Rank 15: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=12|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=14|TP=0|ip-26-0-167-9]: [RECV META] Rank 14: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=13|TP=0|ip-26-0-167-9]: [RECV META] Rank 13: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=5|TP=0|ip-26-0-164-187]: [RECV META] Rank 5: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=15|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=14|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=13|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=5|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=15|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=14|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=13|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=5|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=13|TP=0|ip-26-0-167-9]: [RECV META] Rank 13: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=13|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=13|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=6|TP=0|ip-26-0-164-187]: [RECV META] Rank 6: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=6|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=12|TP=0|ip-26-0-167-9]: [RECV META] Rank 12: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=12|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=12|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=6|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=15|TP=0|ip-26-0-167-9]: [RECV META] Rank 15: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=15|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=15|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=7|TP=0|ip-26-0-164-187]: [RECV META] Rank 7: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=4|TP=0|ip-26-0-164-187]: [RECV META] Rank 4: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=7|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=7|TP=0|ip-26-0-164-187]: [RECV META] Rank 7: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=4|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=7|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=9|TP=0|ip-26-0-166-244]: [RECV META] Rank 9: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=9|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=9|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=4|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=7|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=7|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=6|TP=0|ip-26-0-164-187]: [RECV META] Rank 6: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=6|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=6|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=10|TP=0|ip-26-0-166-244]: [RECV META] Rank 10: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=8|TP=0|ip-26-0-166-244]: [RECV META] Rank 8: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=8|TP=0|ip-26-0-166-244]: [RECV META] Rank 8: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=10|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=8|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=8|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=10|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=8|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=8|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=9|TP=0|ip-26-0-166-244]: [RECV META] Rank 9: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=10|TP=0|ip-26-0-166-244]: [RECV META] Rank 10: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=9|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=9|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=10|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=10|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=11|TP=0|ip-26-0-166-244]: [RECV META] Rank 11: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=11|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=11|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=4|TP=0|ip-26-0-164-187]: [RECV META] Rank 4: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=4|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=4|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=5|TP=0|ip-26-0-164-187]: [RECV META] Rank 5: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=2|TP=0|ip-26-0-163-43]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=2|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=2|TP=0|ip-26-0-163-43]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=5|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=2|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=2|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=2|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=5|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=11|TP=0|ip-26-0-166-244]: [RECV META] Rank 11: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=1|TP=0|ip-26-0-163-43]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=3|TP=0|ip-26-0-163-43]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=1|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=11|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=3|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=11|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=1|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=3|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=3|TP=0|ip-26-0-163-43]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=3|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=3|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=1|TP=0|ip-26-0-163-43]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=1|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=1|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=14|TP=0|ip-26-0-167-9]: [RECV META] Rank 14: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=14|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=12|TP=0|ip-26-0-167-9]: [RECV META] Rank 12: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=14|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=12|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=12|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=15|TP=0|ip-26-0-167-9]: [RECV META] Rank 15: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=14|TP=0|ip-26-0-167-9]: [RECV META] Rank 14: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=5|TP=0|ip-26-0-164-187]: [RECV META] Rank 5: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=13|TP=0|ip-26-0-167-9]: [RECV META] Rank 13: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=15|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=14|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=13|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=15|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=14|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=13|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=13|TP=0|ip-26-0-167-9]: [RECV META] Rank 13: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=5|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=13|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=5|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=13|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=6|TP=0|ip-26-0-164-187]: [RECV META] Rank 6: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=12|TP=0|ip-26-0-167-9]: [RECV META] Rank 12: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=6|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=12|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=6|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=12|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=15|TP=0|ip-26-0-167-9]: [RECV META] Rank 15: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=15|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=15|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=7|TP=0|ip-26-0-164-187]: [RECV META] Rank 7: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=4|TP=0|ip-26-0-164-187]: [RECV META] Rank 4: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=7|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=4|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=7|TP=0|ip-26-0-164-187]: [RECV META] Rank 7: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=7|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=4|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=7|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=7|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=9|TP=0|ip-26-0-166-244]: [RECV META] Rank 9: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=6|TP=0|ip-26-0-164-187]: [RECV META] Rank 6: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=9|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=9|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=6|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=6|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=10|TP=0|ip-26-0-166-244]: [RECV META] Rank 10: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=8|TP=0|ip-26-0-166-244]: [RECV META] Rank 8: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=8|TP=0|ip-26-0-166-244]: [RECV META] Rank 8: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=10|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=8|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=8|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=10|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=8|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=8|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=9|TP=0|ip-26-0-166-244]: [RECV META] Rank 9: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=9|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=10|TP=0|ip-26-0-166-244]: [RECV META] Rank 10: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=9|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=10|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=10|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=11|TP=0|ip-26-0-166-244]: [RECV META] Rank 11: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=11|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=11|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=4|TP=0|ip-26-0-164-187]: [RECV META] Rank 4: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=4|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=4|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=2|TP=0|ip-26-0-163-43]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=2|TP=0|ip-26-0-163-43]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=2|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=2|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=2|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=5|TP=0|ip-26-0-164-187]: [RECV META] Rank 5: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=2|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=5|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=11|TP=0|ip-26-0-166-244]: [RECV META] Rank 11: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=5|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=3|TP=0|ip-26-0-163-43]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=1|TP=0|ip-26-0-163-43]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=3|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=1|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=3|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=1|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=11|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=11|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=3|TP=0|ip-26-0-163-43]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=3|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=3|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=1|TP=0|ip-26-0-163-43]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=1|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=1|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=14|TP=0|ip-26-0-167-9]: [RECV META] Rank 14: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=12|TP=0|ip-26-0-167-9]: [RECV META] Rank 12: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=14|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=12|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=14|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=12|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=14|TP=0|ip-26-0-167-9]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=1|PP=12|TP=0|ip-26-0-167-9]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=1|PP=14|TP=0|ip-26-0-167-9]: [RECV META] Rank 14: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=15|TP=0|ip-26-0-167-9]: [RECV META] Rank 15: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=13|TP=0|ip-26-0-167-9]: [RECV META] Rank 13: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=14|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 13: [14] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '13:14', but store->get('13:14') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=15|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=13|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=14|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=15|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=13|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=14|TP=0|ip-26-0-167-9]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=1|PP=5|TP=0|ip-26-0-164-187]: [RECV META] Rank 5: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=15|TP=0|ip-26-0-167-9]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=0|PP=13|TP=0|ip-26-0-167-9]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [WARNING|DP=1|PP=5|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=13|TP=0|ip-26-0-167-9]: [RECV META] Rank 13: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=5|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=13|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 12: [13] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '12:13', but store->get('12:13') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=5|TP=0|ip-26-0-164-187]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [WARNING|DP=1|PP=13|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=12|TP=0|ip-26-0-167-9]: [RECV META] Rank 12: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=13|TP=0|ip-26-0-167-9]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [WARNING|DP=0|PP=12|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 11: [12] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '11:12', but store->get('11:12') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=6|TP=0|ip-26-0-164-187]: [RECV META] Rank 6: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=6|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=6|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=12|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=12|TP=0|ip-26-0-167-9]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=1|PP=15|TP=0|ip-26-0-167-9]: [RECV META] Rank 15: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=15|TP=0|ip-26-0-167-9]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 14: [15] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '14:15', but store->get('14:15') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=15|TP=0|ip-26-0-167-9]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=15|TP=0|ip-26-0-167-9]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=0|PP=6|TP=0|ip-26-0-164-187]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=1|PP=4|TP=0|ip-26-0-164-187]: [RECV META] Rank 4: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=7|TP=0|ip-26-0-164-187]: [RECV META] Rank 7: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=4|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=7|TP=0|ip-26-0-164-187]: [RECV META] Rank 7: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=4|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=7|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=7|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: [7] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '6:7', but store->get('6:7') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=4|TP=0|ip-26-0-164-187]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [WARNING|DP=0|PP=7|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=7|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=7|TP=0|ip-26-0-164-187]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=1|PP=7|TP=0|ip-26-0-164-187]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=1|PP=9|TP=0|ip-26-0-166-244]: [RECV META] Rank 9: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=9|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=9|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=9|TP=0|ip-26-0-166-244]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=1|PP=6|TP=0|ip-26-0-164-187]: [RECV META] Rank 6: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=6|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: [6] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '5:6', but store->get('5:6') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=6|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=6|TP=0|ip-26-0-164-187]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=0|PP=10|TP=0|ip-26-0-166-244]: [RECV META] Rank 10: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=8|TP=0|ip-26-0-166-244]: [RECV META] Rank 8: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=8|TP=0|ip-26-0-166-244]: [RECV META] Rank 8: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=10|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=8|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=8|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 7: [8] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '7:8', but store->get('7:8') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=10|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=8|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=8|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=10|TP=0|ip-26-0-166-244]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=0|PP=8|TP=0|ip-26-0-166-244]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=0|PP=9|TP=0|ip-26-0-166-244]: [RECV META] Rank 9: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=8|TP=0|ip-26-0-166-244]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [WARNING|DP=0|PP=9|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 8: [9] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '8:9', but store->get('8:9') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=10|TP=0|ip-26-0-166-244]: [RECV META] Rank 10: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=9|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=1|PP=10|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 9: [10] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '9:10', but store->get('9:10') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=9|TP=0|ip-26-0-166-244]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [WARNING|DP=1|PP=10|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=11|TP=0|ip-26-0-166-244]: [RECV META] Rank 11: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=1|PP=10|TP=0|ip-26-0-166-244]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [WARNING|DP=1|PP=11|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=11|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=11|TP=0|ip-26-0-166-244]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=0|PP=4|TP=0|ip-26-0-164-187]: [RECV META] Rank 4: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=4|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: [4] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '3:4', but store->get('3:4') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=4|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=4|TP=0|ip-26-0-164-187]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=1|PP=2|TP=0|ip-26-0-163-43]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=2|TP=0|ip-26-0-163-43]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=2|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=2|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=2|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=2|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=2|TP=0|ip-26-0-163-43]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=0|PP=5|TP=0|ip-26-0-164-187]: [RECV META] Rank 5: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=2|TP=0|ip-26-0-163-43]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=0|PP=11|TP=0|ip-26-0-166-244]: [RECV META] Rank 11: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=1|TP=0|ip-26-0-163-43]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=3|TP=0|ip-26-0-163-43]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=5|TP=0|ip-26-0-164-187]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: [5] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '4:5', but store->get('4:5') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=1|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=3|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=0|PP=5|TP=0|ip-26-0-164-187]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=1|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=3|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [WARNING|DP=0|PP=11|TP=0|ip-26-0-166-244]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 10: [11] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '10:11', but store->get('10:11') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [ERROR|DP=0|PP=5|TP=0|ip-26-0-164-187]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=0|PP=1|TP=0|ip-26-0-163-43]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=0|PP=3|TP=0|ip-26-0-163-43]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [WARNING|DP=0|PP=11|TP=0|ip-26-0-166-244]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=0|PP=11|TP=0|ip-26-0-166-244]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=1|PP=3|TP=0|ip-26-0-163-43]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=3|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=3|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=3|TP=0|ip-26-0-163-43]: Max retries reached, giving up on communication +12/27/2024 17:35:39 [ERROR|DP=1|PP=1|TP=0|ip-26-0-163-43]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=1|TP=0|ip-26-0-163-43]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:35:39 [WARNING|DP=1|PP=1|TP=0|ip-26-0-163-43]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:35:39 [ERROR|DP=1|PP=1|TP=0|ip-26-0-163-43]: Max retries reached, giving up on communication +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs)trainer.train(dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + trainer.train(dataloader)outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + trainer.train(dataloader) + trainer.train(dataloader) + trainer.train(dataloader) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return f(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + output = model(**micro_batch) + output = model(**micro_batch) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return f(*args, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = model(**micro_batch) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +return self._call_impl(*args, **kwargs) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + +return f(*args, **kwargs) +return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return self._call_impl(*args, **kwargs) +trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs = self.pipeline_engine.train_batch_iter(outputs = self.pipeline_engine.train_batch_iter( + + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) +return forward_call(*args, **kwargs) + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model)return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward +return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +return self._call_impl(*args, **kwargs) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + sharded_logits = self.model( + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)outputs = self.pipeline_engine.train_batch_iter(outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return self._call_impl(*args, **kwargs) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + sharded_logits = self.model( + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return forward_call(*args, **kwargs) +return self._call_impl(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + +return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + output = model(**micro_batch)output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward +output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) +hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states)hidden_encoder_states = encoder_block(**hidden_encoder_states) + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + hidden_encoder_states = encoder_block(**hidden_encoder_states) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +hidden_encoder_states = encoder_block(**hidden_encoder_states) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + hidden_encoder_states = encoder_block(**hidden_encoder_states) + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return self._call_impl(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return forward_call(*args, **kwargs) + + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward +return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + +return forward_call(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + +return forward_call(*args, **kwargs) + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + + sharded_logits = self.model( + sharded_logits = self.model( File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + new_kwargs[name] = recv_from_pipeline_state_buffer( +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + activation = pipeline_state.activations_buffer.popleft() + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward +IndexError: pop from an empty deque + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + activation = pipeline_state.activations_buffer.popleft() + return self._call_impl(*args, **kwargs)IndexError +: pop from an empty deque + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + new_kwargs[name] = recv_from_pipeline_state_buffer( + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + return forward_call(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + sharded_logits = self.model( + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer +new_kwargs[name] = recv_from_pipeline_state_buffer( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0]return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + new_kwargs[name] = recv_from_pipeline_state_buffer( +new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer +new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0]return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states)hidden_encoder_states = encoder_block(**hidden_encoder_states) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + activation = pipeline_state.activations_buffer.popleft() + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward +IndexError: pop from an empty deque + activation = pipeline_state.activations_buffer.popleft() +IndexError : activation = pipeline_state.activations_buffer.popleft() +pop from an empty deque +IndexError: pop from an empty deque + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + new_kwargs[name] = recv_from_pipeline_state_buffer(new_kwargs[name] = recv_from_pipeline_state_buffer( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + hidden_encoder_states = encoder_block(**hidden_encoder_states) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + hidden_encoder_states = encoder_block(**hidden_encoder_states)hidden_encoder_states = encoder_block(**hidden_encoder_states)hidden_encoder_states = encoder_block(**hidden_encoder_states) + + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) return self._call_impl(*args, **kwargs) + +return forward_call(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + activation = pipeline_state.activations_buffer.popleft()IndexError + : activation = pipeline_state.activations_buffer.popleft()pop from an empty dequeIndexError + +: pop from an empty deque +IndexError: pop from an empty deque + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + activation = pipeline_state.activations_buffer.popleft() + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer +IndexError: pop from an empty deque + activation = pipeline_state.activations_buffer.popleft() + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward +new_kwargs[name] = recv_from_pipeline_state_buffer( + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +IndexError: pop from an empty deque + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) +IndexError: pop from an empty deque + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward +IndexError: pop from an empty deque + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + activation = pipeline_state.activations_buffer.popleft() + hidden_encoder_states = encoder_block(**hidden_encoder_states)IndexError +: pop from an empty deque + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer +new_kwargs[name] = recv_from_pipeline_state_buffer( + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty dequeactivation = pipeline_state.activations_buffer.popleft()activation = pipeline_state.activations_buffer.popleft() + + +IndexErrorIndexError: : pop from an empty dequepop from an empty deque + + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque +[2024-12-27 17:35:59,809] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1070838 closing signal SIGTERM +[2024-12-27 17:35:59,809] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1070840 closing signal SIGTERM +[2024-12-27 17:35:59,809] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1070841 closing signal SIGTERM +[2024-12-27 17:35:59,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1070842 closing signal SIGTERM +[2024-12-27 17:35:59,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1070843 closing signal SIGTERM +[2024-12-27 17:35:59,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1070844 closing signal SIGTERM +[2024-12-27 17:35:59,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2110732 closing signal SIGTERM +[2024-12-27 17:35:59,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2110733 closing signal SIGTERM +[2024-12-27 17:35:59,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2110734 closing signal SIGTERM +[2024-12-27 17:35:59,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2110735 closing signal SIGTERM +[2024-12-27 17:35:59,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2110736 closing signal SIGTERM +[2024-12-27 17:35:59,815] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2110737 closing signal SIGTERM +[2024-12-27 17:35:59,821] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2656308 closing signal SIGTERM +[2024-12-27 17:35:59,821] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2656309 closing signal SIGTERM +[2024-12-27 17:35:59,821] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2656310 closing signal SIGTERM +[2024-12-27 17:35:59,821] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2656311 closing signal SIGTERM +[2024-12-27 17:35:59,821] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2656312 closing signal SIGTERM +[2024-12-27 17:35:59,822] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2656313 closing signal SIGTERM +[2024-12-27 17:35:59,827] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1074706 closing signal SIGTERM +[2024-12-27 17:35:59,827] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1074707 closing signal SIGTERM +[2024-12-27 17:35:59,827] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1074709 closing signal SIGTERM +[2024-12-27 17:35:59,827] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1074710 closing signal SIGTERM +[2024-12-27 17:35:59,827] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1074711 closing signal SIGTERM +[2024-12-27 17:36:01,644] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1074705) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 17:36:01,686] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_q_bstmsz/13803408_ohopoal3/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:35:39 + host : ip-26-0-166-244.ec2.internal + rank : 19 (local_rank: 3) + exitcode : 1 (pid: 1074708) + error_file: /tmp/torchelastic_q_bstmsz/13803408_ohopoal3/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +[2]: + time : 2024-12-27_17:35:39 + host : ip-26-0-166-244.ec2.internal + rank : 23 (local_rank: 7) + exitcode : 1 (pid: 1074712) + error_file: /tmp/torchelastic_q_bstmsz/13803408_ohopoal3/attempt_0/7/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:35:39 + host : ip-26-0-166-244.ec2.internal + rank : 16 (local_rank: 0) + exitcode : 1 (pid: 1074705) + error_file: /tmp/torchelastic_q_bstmsz/13803408_ohopoal3/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +============================================================ +[2024-12-27 17:36:01,890] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1070837) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 17:36:01,928] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_te0dow0k/13803408_pcsup1_7/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:35:39 + host : ip-26-0-167-9.ec2.internal + rank : 26 (local_rank: 2) + exitcode : 1 (pid: 1070839) + error_file: /tmp/torchelastic_te0dow0k/13803408_pcsup1_7/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:35:39 + host : ip-26-0-167-9.ec2.internal + rank : 24 (local_rank: 0) + exitcode : 1 (pid: 1070837) + error_file: /tmp/torchelastic_te0dow0k/13803408_pcsup1_7/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +============================================================ +[2024-12-27 17:36:02,032] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2110730) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 17:36:02,073] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_s03lfupi/13803408_0bwwxs1l/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:35:39 + host : ip-26-0-164-187.ec2.internal + rank : 9 (local_rank: 1) + exitcode : 1 (pid: 2110731) + error_file: /tmp/torchelastic_s03lfupi/13803408_0bwwxs1l/attempt_0/1/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:35:39 + host : ip-26-0-164-187.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 2110730) + error_file: /tmp/torchelastic_s03lfupi/13803408_0bwwxs1l/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + IndexError: pop from an empty deque + +============================================================ +srun: error: ip-26-0-166-244: task 2: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13803408.0 +slurmstepd: error: *** STEP 13803408.0 ON ip-26-0-163-43 CANCELLED AT 2024-12-27T17:36:02 *** +[2024-12-27 17:36:02,185] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-27 17:36:02,185] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2656310 closing signal SIGTERM +[2024-12-27 17:36:02,186] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2656313 closing signal SIGTERM +srun: error: ip-26-0-167-9: task 3: Terminated +srun: error: ip-26-0-164-187: task 1: Terminated +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2656232 got signal: 15 +srun: error: ip-26-0-163-43: task 0: Exited with exit code 1 +srun: Force Terminated StepId=13803408.0 diff --git a/logs/13803433-bench_469G_dp2_tp2_pp8_acc128_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13803433-bench_469G_dp2_tp2_pp8_acc128_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..bbdf4720fbc8dcbf09ff4ccb7a83ac89354af80f --- /dev/null +++ b/logs/13803433-bench_469G_dp2_tp2_pp8_acc128_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,8095 @@ ++ '[' -z 13803433 ']' ++ unset FI_PROVIDER ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames ip-26-0-166-214,ip-26-0-168-95,ip-26-0-174-196,ip-26-0-175-132 ++ export 'NODELIST=ip-26-0-166-214 +ip-26-0-168-95 +ip-26-0-174-196 +ip-26-0-175-132' ++ NODELIST='ip-26-0-166-214 +ip-26-0-168-95 +ip-26-0-174-196 +ip-26-0-175-132' +++ scontrol show hostnames ip-26-0-166-214,ip-26-0-168-95,ip-26-0-174-196,ip-26-0-175-132 +++ head -n1 ++ export MASTER_NODE=ip-26-0-166-214 ++ MASTER_NODE=ip-26-0-166-214 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ export NCCL_P2P_LEVEL=LOC ++ NCCL_P2P_LEVEL=LOC ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-166-214' +Master node: ip-26-0-166-214 ++ echo 'All nodes: ip-26-0-166-214 +ip-26-0-168-95 +ip-26-0-174-196 +ip-26-0-175-132' +All nodes: ip-26-0-166-214 +ip-26-0-168-95 +ip-26-0-174-196 +ip-26-0-175-132 ++ echo 'World size: 32' +World size: 32 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13803433 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-166-214:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_469G_dp2_tp2_pp8_acc128_mbs1_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-27 17:39:45,113] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:39:45,135] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:39:45,113] torch.distributed.run: [WARNING] +[2024-12-27 17:39:45,113] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:39:45,113] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:39:45,113] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:39:45,175] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:39:45,135] torch.distributed.run: [WARNING] +[2024-12-27 17:39:45,135] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:39:45,135] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:39:45,135] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:39:45,176] torch.distributed.run: [WARNING] +[2024-12-27 17:39:45,176] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:39:45,176] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:39:45,176] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:39:45,316] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:39:45,316] torch.distributed.run: [WARNING] +[2024-12-27 17:39:45,316] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:39:45,316] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:39:45,316] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-166-214:1132723:1132723 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1132723:1132723 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1132723:1132723 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1132723:1132723 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-214:1132725:1132725 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1132724:1132724 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-175-132:679635:679635 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-175-132:679632:679632 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-175-132:679633:679633 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-175-132:679639:679639 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-175-132:679638:679638 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-175-132:679636:679636 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-175-132:679634:679634 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-175-132:679637:679637 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1132727:1132727 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1132725:1132725 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1132724:1132724 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-168-95:1632401:1632401 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1132727:1132727 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1132728:1132728 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1132725:1132725 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1132724:1132724 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1132725:1132725 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1132724:1132724 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1132726:1132726 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1132727:1132727 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1132727:1132727 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1132730:1132730 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1632407:1632407 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1632408:1632408 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1632405:1632405 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1632404:1632404 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-196:3624155:3624155 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-196:3624149:3624149 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-196:3624151:3624151 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-196:3624156:3624156 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1632406:1632406 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-196:3624154:3624154 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-196:3624152:3624152 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1132729:1132729 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-196:3624150:3624150 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-174-196:3624153:3624153 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1632403:1632403 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1632402:1632402 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1132728:1132728 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1132726:1132726 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1132728:1132728 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1132728:1132728 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1132730:1132730 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1132729:1132729 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1132726:1132726 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1132726:1132726 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1632401:1632401 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-166-214:1132729:1132729 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1132729:1132729 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1132730:1132730 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1132730:1132730 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1632401:1632401 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1632401:1632401 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-175-132:679632:679632 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.175.132<0> +ip-26-0-175-132:679635:679635 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.175.132<0> +ip-26-0-175-132:679637:679637 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.175.132<0> +ip-26-0-175-132:679639:679639 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.175.132<0> +ip-26-0-175-132:679636:679636 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.175.132<0> +ip-26-0-175-132:679633:679633 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.175.132<0> +ip-26-0-175-132:679634:679634 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.175.132<0> +ip-26-0-175-132:679638:679638 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.175.132<0> +ip-26-0-175-132:679632:679632 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-175-132:679637:679637 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-175-132:679637:679637 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-175-132:679632:679632 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-175-132:679635:679635 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-175-132:679635:679635 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-175-132:679639:679639 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-175-132:679639:679639 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-175-132:679634:679634 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-175-132:679636:679636 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-175-132:679638:679638 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-175-132:679634:679634 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-175-132:679636:679636 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-175-132:679638:679638 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-175-132:679633:679633 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-175-132:679633:679633 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1632407:1632407 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1632408:1632408 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1632405:1632405 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1632404:1632404 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1632406:1632406 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1632403:1632403 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1632402:1632402 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1632408:1632408 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1632408:1632408 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1632407:1632407 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1632407:1632407 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1632405:1632405 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1632405:1632405 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1632404:1632404 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1632404:1632404 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1632403:1632403 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1632403:1632403 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1632406:1632406 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1632406:1632406 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1632402:1632402 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1632402:1632402 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-196:3624155:3624155 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.174.196<0> +ip-26-0-174-196:3624149:3624149 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.174.196<0> +ip-26-0-174-196:3624151:3624151 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.174.196<0> +ip-26-0-174-196:3624154:3624154 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.174.196<0> +ip-26-0-174-196:3624156:3624156 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.174.196<0> +ip-26-0-174-196:3624152:3624152 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.174.196<0> +ip-26-0-174-196:3624150:3624150 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.174.196<0> +ip-26-0-174-196:3624153:3624153 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.174.196<0> +ip-26-0-174-196:3624149:3624149 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-196:3624154:3624154 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-196:3624154:3624154 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-196:3624149:3624149 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-196:3624155:3624155 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-196:3624155:3624155 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-196:3624153:3624153 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-196:3624156:3624156 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-196:3624153:3624153 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-196:3624156:3624156 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-196:3624150:3624150 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-196:3624150:3624150 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-196:3624151:3624151 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-196:3624151:3624151 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-174-196:3624152:3624152 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-174-196:3624152:3624152 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/253 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Using network Libfabric +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-175-132:679639:679795 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-175-132:679639:679795 [7] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-175-132:679635:679793 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-175-132:679635:679793 [3] NCCL INFO Using network Libfabric +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-175-132:679633:679794 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-175-132:679633:679794 [1] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Using network Libfabric +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-175-132:679637:679798 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-175-132:679637:679798 [5] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Using network Libfabric +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-175-132:679636:679797 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-175-132:679636:679797 [4] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Using network Libfabric +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-175-132:679638:679800 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-175-132:679638:679800 [6] NCCL INFO Using network Libfabric +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-175-132:679634:679796 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-175-132:679634:679796 [2] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Using network Libfabric +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-175-132:679632:679799 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-175-132:679632:679799 [0] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO comm 0x885a7b0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO comm 0x8946800 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO comm 0xa3c6e80 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-175-132:679632:679799 [0] NCCL INFO comm 0x9327030 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-175-132:679633:679794 [1] NCCL INFO comm 0x8fd6010 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO comm 0x88e5f40 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO comm 0x8f470f0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO comm 0x914b280 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO comm 0x8fa5550 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-175-132:679634:679796 [2] NCCL INFO comm 0x88cc240 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-175-132:679635:679793 [3] NCCL INFO comm 0x8ec0580 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-175-132:679636:679797 [4] NCCL INFO comm 0x91b7dd0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-175-132:679638:679800 [6] NCCL INFO comm 0xa282f90 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-175-132:679639:679795 [7] NCCL INFO comm 0x9111de0 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO comm 0x90273c0 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO comm 0xa0a6620 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-175-132:679637:679798 [5] NCCL INFO comm 0xa28efa0 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO comm 0x9874770 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO comm 0x973c050 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO comm 0x8fafa70 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO comm 0x90c65a0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO comm 0x9089520 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO comm 0x9e91460 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO comm 0x92e3280 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO comm 0xa1debe0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO comm 0x84966b0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO comm 0x89644f0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO comm 0x984ff90 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO comm 0x8a56e80 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO comm 0x8af6540 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO comm 0x951e030 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO comm 0x8c88e00 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf329f66c66fd73c0 - Init START +ip-26-0-175-132:679639:679795 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679636:679797 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679635:679793 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679634:679796 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679632:679799 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-175-132:679632:679799 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-175-132:679636:679797 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-175-132:679636:679797 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-175-132:679638:679800 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-175-132:679638:679800 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-175-132:679637:679798 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-175-132:679637:679798 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-175-132:679634:679796 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-175-132:679634:679796 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-175-132:679635:679793 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-175-132:679635:679793 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-175-132:679639:679795 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-175-132:679639:679795 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-175-132:679633:679794 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-175-132:679633:679794 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-175-132:679639:679795 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 15/-1/-1->31->-1 [3] 15/-1/-1->31->-1 +ip-26-0-175-132:679639:679795 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Trees [0] 16/-1/-1->0->-1 [1] 16/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Trees [0] 8/24/-1->16->0 [1] 8/24/-1->16->0 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-175-132:679638:679800 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-175-132:679638:679800 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679636:679797 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679636:679797 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679637:679798 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-175-132:679637:679798 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679633:679794 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-175-132:679633:679794 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679635:679793 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-175-132:679632:679799 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-175-132:679635:679793 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679632:679799 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679634:679796 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-175-132:679634:679796 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679639:679795 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-175-132:679638:679800 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-175-132:679636:679797 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-175-132:679633:679794 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-175-132:679635:679793 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-175-132:679632:679799 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-175-132:679637:679798 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-175-132:679634:679796 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Connected all rings +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Connected all rings +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Connected all rings +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Connected all rings +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Connected all rings +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Connected all rings +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Connected all rings +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Connected all rings +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Connected all rings +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Connected all rings +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Connected all rings +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Connected all rings +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Connected all rings +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Connected all rings +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Connected all rings +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Connected all rings +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Connected all rings +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Connected all rings +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Connected all rings +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Connected all rings +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:679799 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:679797 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679638:679800 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:679796 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679635:679793 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:679798 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:679794 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679639:679795 [7] NCCL INFO Connected all trees +ip-26-0-175-132:679639:679795 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679639:679795 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679639:679795 [7] NCCL INFO comm 0x9111de0 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-166-214:1132723:1132886 [0] NCCL INFO comm 0x90273c0 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO Connected all trees +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO Connected all trees +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679634:679796 [2] NCCL INFO Connected all trees +ip-26-0-175-132:679634:679796 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679634:679796 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO Connected all trees +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624155:3624314 [6] NCCL INFO comm 0xa3c6e80 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-174-196:3624151:3624311 [2] NCCL INFO comm 0x88e5f40 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO Connected all trees +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679634:679796 [2] NCCL INFO comm 0x88cc240 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624150:3624317 [1] NCCL INFO comm 0x8fa5550 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-175-132:679638:679800 [6] NCCL INFO Connected all trees +ip-26-0-175-132:679638:679800 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679638:679800 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679637:679798 [5] NCCL INFO Connected all trees +ip-26-0-175-132:679637:679798 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679637:679798 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679633:679794 [1] NCCL INFO Connected all trees +ip-26-0-175-132:679633:679794 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679633:679794 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624154:3624315 [5] NCCL INFO comm 0x8946800 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO Connected all trees +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132729:1132892 [6] NCCL INFO comm 0x90c65a0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-175-132:679638:679800 [6] NCCL INFO comm 0xa282f90 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-175-132:679637:679798 [5] NCCL INFO comm 0xa28efa0 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-175-132:679633:679794 [1] NCCL INFO comm 0x8fd6010 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-166-214:1132725:1132890 [2] NCCL INFO comm 0x8fafa70 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-166-214:1132724:1132889 [1] NCCL INFO comm 0xa0a6620 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO Connected all trees +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132728:1132888 [5] NCCL INFO comm 0x9e91460 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-168-95:1632403:1632561 [2] NCCL INFO comm 0x84966b0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO Connected all trees +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632402:1632564 [1] NCCL INFO comm 0x92e3280 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-168-95:1632407:1632568 [6] NCCL INFO comm 0x951e030 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO Connected all trees +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO Connected all trees +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO Connected all trees +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632406:1632563 [5] NCCL INFO comm 0x8a56e80 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-175-132:679635:679793 [3] NCCL INFO Connected all trees +ip-26-0-175-132:679635:679793 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679635:679793 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624156:3624316 [7] NCCL INFO comm 0x885a7b0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624152:3624312 [3] NCCL INFO comm 0x914b280 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-175-132:679635:679793 [3] NCCL INFO comm 0x8ec0580 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-175-132:679632:679799 [0] NCCL INFO Connected all trees +ip-26-0-175-132:679632:679799 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679632:679799 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO Connected all trees +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO Connected all trees +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679636:679797 [4] NCCL INFO Connected all trees +ip-26-0-175-132:679636:679797 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679636:679797 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132730:1132893 [7] NCCL INFO comm 0x9089520 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-175-132:679632:679799 [0] NCCL INFO comm 0x9327030 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-166-214:1132727:1132887 [4] NCCL INFO comm 0x973c050 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-175-132:679636:679797 [4] NCCL INFO comm 0x91b7dd0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-174-196:3624149:3624313 [0] NCCL INFO comm 0x8af6540 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-174-196:3624153:3624310 [4] NCCL INFO comm 0x8f470f0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-166-214:1132726:1132891 [3] NCCL INFO comm 0x9874770 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO Connected all trees +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO Connected all trees +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632404:1632567 [3] NCCL INFO comm 0x984ff90 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-168-95:1632408:1632566 [7] NCCL INFO comm 0x8c88e00 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO Connected all trees +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO Connected all trees +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632405:1632565 [4] NCCL INFO comm 0x89644f0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-168-95:1632401:1632562 [0] NCCL INFO comm 0xa1debe0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf329f66c66fd73c0 - Init COMPLETE +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Using network Libfabric +ip-26-0-175-132:679636:680042 [4] NCCL INFO Using network Libfabric +ip-26-0-175-132:679639:680040 [7] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Using network Libfabric +ip-26-0-175-132:679635:680045 [3] NCCL INFO Using network Libfabric +ip-26-0-175-132:679634:680046 [2] NCCL INFO Using network Libfabric +ip-26-0-175-132:679632:680039 [0] NCCL INFO Using network Libfabric +ip-26-0-175-132:679638:680044 [6] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Using network Libfabric +ip-26-0-175-132:679633:680043 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Using network Libfabric +ip-26-0-175-132:679637:680041 [5] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Using network Libfabric +ip-26-0-175-132:679639:680040 [7] NCCL INFO comm 0x934f000 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO comm 0xa2e3420 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO comm 0x926a8c0 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-175-132:679638:680044 [6] NCCL INFO comm 0xa4be5f0 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO comm 0x8a95130 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO comm 0xa6035c0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-175-132:679637:680041 [5] NCCL INFO comm 0xa4cb8c0 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-175-132:679636:680042 [4] NCCL INFO comm 0x93f1ee0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-175-132:679635:680045 [3] NCCL INFO comm 0x90fdac0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-175-132:679634:680046 [2] NCCL INFO comm 0x8b08640 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-175-132:679633:680043 [1] NCCL INFO comm 0x9212a80 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-175-132:679632:680039 [0] NCCL INFO comm 0x9563d00 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO comm 0x8b817a0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO comm 0x8ec5880 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO comm 0x9182600 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO comm 0x8b221c0 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO comm 0x9388300 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO comm 0x91e2140 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO comm 0x8d30760 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO comm 0x9758700 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO comm 0x92c3420 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO comm 0x9300970 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO comm 0xa0cb720 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO comm 0x8c91630 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO comm 0x8ba0d60 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO comm 0x99789c0 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO comm 0x91eafb0 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO comm 0x9aaf1b0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO comm 0x9a8b170 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO comm 0x86d0c70 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO comm 0x951d2c0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO comm 0xa419040 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9dbdbcbac26299e4 - Init START +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-175-132:679639:680040 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679638:680044 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679636:680042 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679637:680041 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679635:680045 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679634:680046 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679633:680043 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679632:680039 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-175-132:679635:680045 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-175-132:679633:680043 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-175-132:679636:680042 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-175-132:679634:680046 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-175-132:679632:680039 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-175-132:679637:680041 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-175-132:679639:680040 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-175-132:679638:680044 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679639:680040 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 15/-1/-1->31->-1 [3] 15/-1/-1->31->-1 +ip-26-0-175-132:679639:680040 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679634:680046 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-175-132:679638:680044 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-175-132:679634:680046 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679638:680044 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Trees [0] 16/-1/-1->0->-1 [1] 16/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679636:680042 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-175-132:679632:680039 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-175-132:679636:680042 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679632:680039 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679633:680043 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-175-132:679635:680045 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679633:680043 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679635:680045 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679637:680041 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-175-132:679637:680041 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Trees [0] 8/24/-1->16->0 [1] 8/24/-1->16->0 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Connected all rings +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Connected all rings +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Connected all rings +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Connected all rings +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Connected all rings +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Connected all rings +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Connected all rings +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Connected all rings +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Connected all rings +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Connected all rings +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Connected all rings +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Connected all rings +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Connected all rings +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Connected all rings +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Connected all rings +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Connected all rings +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Connected all rings +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Connected all rings +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Connected all rings +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Connected all rings +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Connected all rings +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Connected all rings +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Connected all rings +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Connected all rings +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Connected all rings +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680039 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680045 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680041 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679633:680043 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679636:680042 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679638:680044 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680046 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679639:680040 [7] NCCL INFO Connected all trees +ip-26-0-175-132:679639:680040 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679639:680040 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132723:1133122 [0] NCCL INFO comm 0x926a8c0 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-175-132:679639:680040 [7] NCCL INFO comm 0x934f000 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO Connected all trees +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624151:3624549 [2] NCCL INFO comm 0x8b221c0 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132725:1133126 [2] NCCL INFO comm 0x91eafb0 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-175-132:679634:680046 [2] NCCL INFO Connected all trees +ip-26-0-175-132:679634:680046 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679634:680046 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132724:1133124 [1] NCCL INFO comm 0xa2e3420 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO Connected all trees +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679634:680046 [2] NCCL INFO comm 0x8b08640 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO Connected all trees +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO Connected all trees +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO Connected all trees +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132729:1133128 [6] NCCL INFO comm 0x9300970 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-174-196:3624155:3624546 [6] NCCL INFO comm 0xa6035c0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-175-132:679638:680044 [6] NCCL INFO Connected all trees +ip-26-0-175-132:679638:680044 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679638:680044 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632407:1632797 [6] NCCL INFO comm 0x9758700 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-168-95:1632403:1632801 [2] NCCL INFO comm 0x86d0c70 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-174-196:3624154:3624548 [5] NCCL INFO comm 0x8b817a0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO Connected all trees +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679638:680044 [6] NCCL INFO comm 0xa4be5f0 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO Connected all trees +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO Connected all trees +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624150:3624544 [1] NCCL INFO comm 0x91e2140 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO Connected all trees +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO Connected all trees +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624153:3624547 [4] NCCL INFO comm 0x9182600 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-174-196:3624156:3624550 [7] NCCL INFO comm 0x8a95130 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-174-196:3624152:3624545 [3] NCCL INFO comm 0x9388300 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-166-214:1132730:1133129 [7] NCCL INFO comm 0x92c3420 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-174-196:3624149:3624551 [0] NCCL INFO comm 0x8d30760 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132726:1133123 [3] NCCL INFO comm 0x9aaf1b0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-175-132:679637:680041 [5] NCCL INFO Connected all trees +ip-26-0-175-132:679637:680041 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679637:680041 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132728:1133127 [5] NCCL INFO comm 0xa0cb720 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-175-132:679635:680045 [3] NCCL INFO Connected all trees +ip-26-0-175-132:679635:680045 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679635:680045 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO Connected all trees +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132727:1133125 [4] NCCL INFO comm 0x99789c0 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-175-132:679633:680043 [1] NCCL INFO Connected all trees +ip-26-0-175-132:679633:680043 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679633:680043 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679632:680039 [0] NCCL INFO Connected all trees +ip-26-0-175-132:679632:680039 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679632:680039 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679637:680041 [5] NCCL INFO comm 0xa4cb8c0 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO Connected all trees +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679636:680042 [4] NCCL INFO Connected all trees +ip-26-0-175-132:679636:680042 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-175-132:679636:680042 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679635:680045 [3] NCCL INFO comm 0x90fdac0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO Connected all trees +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632408:1632796 [7] NCCL INFO comm 0x8ec5880 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-175-132:679633:680043 [1] NCCL INFO comm 0x9212a80 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-175-132:679632:680039 [0] NCCL INFO comm 0x9563d00 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO Connected all trees +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO Connected all trees +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO Connected all trees +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679636:680042 [4] NCCL INFO comm 0x93f1ee0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-168-95:1632404:1632798 [3] NCCL INFO comm 0x9a8b170 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-168-95:1632402:1632800 [1] NCCL INFO comm 0x951d2c0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-168-95:1632406:1632799 [5] NCCL INFO comm 0x8c91630 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-168-95:1632401:1632802 [0] NCCL INFO comm 0xa419040 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +ip-26-0-168-95:1632405:1632795 [4] NCCL INFO comm 0x8ba0d60 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9dbdbcbac26299e4 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO comm 0x92d7810 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xeebe2b6fcd69f481 - Init START +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO comm 0x998b680 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xeebe2b6fcd69f481 - Init START +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO comm 0x927d580 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xeebe2b6fcd69f481 - Init START +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO comm 0xa0df9f0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeebe2b6fcd69f481 - Init START +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO comm 0x9314aa0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xeebe2b6fcd69f481 - Init START +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO comm 0xa2f60e0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeebe2b6fcd69f481 - Init START +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO comm 0x91fe5e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xeebe2b6fcd69f481 - Init START +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO comm 0x9ac3260 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xeebe2b6fcd69f481 - Init START +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679636:680074 [4] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO comm 0x91f4e00 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x44bc682292fe124c - Init START +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO comm 0x8b95980 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x44bc682292fe124c - Init START +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO comm 0xa616280 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x44bc682292fe124c - Init START +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO comm 0x8aa94a0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x44bc682292fe124c - Init START +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO comm 0x8b35540 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44bc682292fe124c - Init START +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO comm 0x939afc0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44bc682292fe124c - Init START +ip-26-0-175-132:679639:680076 [7] NCCL INFO Using network Libfabric +ip-26-0-175-132:679635:680078 [3] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO comm 0x8d45540 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x44bc682292fe124c - Init START +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO comm 0x9196880 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x44bc682292fe124c - Init START +ip-26-0-175-132:679637:680077 [5] NCCL INFO Using network Libfabric +ip-26-0-175-132:679638:680079 [6] NCCL INFO Using network Libfabric +ip-26-0-175-132:679632:680073 [0] NCCL INFO Using network Libfabric +ip-26-0-175-132:679633:680075 [1] NCCL INFO Using network Libfabric +ip-26-0-175-132:679634:680080 [2] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-175-132:679634:680080 [2] NCCL INFO comm 0x8b1b300 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4df034fc67a447f8 - Init START +ip-26-0-175-132:679639:680076 [7] NCCL INFO comm 0x9361cc0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4df034fc67a447f8 - Init START +ip-26-0-175-132:679632:680073 [0] NCCL INFO comm 0x95769c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4df034fc67a447f8 - Init START +ip-26-0-175-132:679635:680078 [3] NCCL INFO comm 0x9110780 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4df034fc67a447f8 - Init START +ip-26-0-175-132:679633:680075 [1] NCCL INFO comm 0x9225740 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4df034fc67a447f8 - Init START +ip-26-0-175-132:679636:680074 [4] NCCL INFO comm 0x9406140 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4df034fc67a447f8 - Init START +ip-26-0-175-132:679638:680079 [6] NCCL INFO comm 0xa4d1e10 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4df034fc67a447f8 - Init START +ip-26-0-175-132:679637:680077 [5] NCCL INFO comm 0xa4de580 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4df034fc67a447f8 - Init START +ip-26-0-175-132:679639:680076 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679638:680079 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679636:680074 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679637:680077 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679635:680078 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679634:680080 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679633:680075 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679632:680073 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO comm 0x976c220 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x25a5be1d9cb5a76f - Init START +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO comm 0x8ed8540 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x25a5be1d9cb5a76f - Init START +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO comm 0x8ca55c0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x25a5be1d9cb5a76f - Init START +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO comm 0x8bb3a20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x25a5be1d9cb5a76f - Init START +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO comm 0xa42e950 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x25a5be1d9cb5a76f - Init START +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO comm 0x9a9eac0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x25a5be1d9cb5a76f - Init START +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO comm 0x95317c0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x25a5be1d9cb5a76f - Init START +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO comm 0x86e4cd0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x25a5be1d9cb5a76f - Init START +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-175-132:679636:680074 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-175-132:679638:680079 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-175-132:679635:680078 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-175-132:679634:680080 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-175-132:679637:680077 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-175-132:679633:680075 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-175-132:679632:680073 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-175-132:679639:680076 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-175-132:679638:680079 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-175-132:679639:680076 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-175-132:679638:680079 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679639:680076 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679637:680077 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-175-132:679637:680077 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679636:680074 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-175-132:679635:680078 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-175-132:679636:680074 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-175-132:679635:680078 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679634:680080 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-175-132:679634:680080 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-175-132:679633:680075 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-175-132:679633:680075 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679632:680073 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-175-132:679632:680073 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679639:680076 [7] NCCL INFO Connected all rings +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Connected all rings +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Connected all rings +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Connected all rings +ip-26-0-175-132:679632:680073 [0] NCCL INFO Connected all rings +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Connected all rings +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Connected all rings +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680073 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679636:680074 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Connected all rings +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:680078 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680077 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680075 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679639:680076 [7] NCCL INFO Connected all trees +ip-26-0-175-132:679639:680076 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679639:680076 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679639:680076 [7] NCCL INFO comm 0x9361cc0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4df034fc67a447f8 - Init COMPLETE +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132730:1133163 [7] NCCL INFO comm 0x92d7810 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xeebe2b6fcd69f481 - Init COMPLETE +ip-26-0-175-132:679632:680073 [0] NCCL INFO Connected all trees +ip-26-0-175-132:679632:680073 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679632:680073 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679632:680073 [0] NCCL INFO comm 0x95769c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4df034fc67a447f8 - Init COMPLETE +ip-26-0-175-132:679633:680075 [1] NCCL INFO Connected all trees +ip-26-0-175-132:679633:680075 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679633:680075 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679633:680075 [1] NCCL INFO comm 0x9225740 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4df034fc67a447f8 - Init COMPLETE +ip-26-0-175-132:679638:680079 [6] NCCL INFO Connected all trees +ip-26-0-175-132:679638:680079 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679638:680079 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679634:680080 [2] NCCL INFO Connected all trees +ip-26-0-175-132:679634:680080 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679634:680080 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Connected all rings +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679638:680079 [6] NCCL INFO comm 0xa4d1e10 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4df034fc67a447f8 - Init COMPLETE +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Connected all rings +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679634:680080 [2] NCCL INFO comm 0x8b1b300 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4df034fc67a447f8 - Init COMPLETE +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Connected all rings +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Connected all rings +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Connected all rings +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Connected all rings +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Connected all rings +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132723:1133157 [0] NCCL INFO comm 0x927d580 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xeebe2b6fcd69f481 - Init COMPLETE +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Connected all rings +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679635:680078 [3] NCCL INFO Connected all trees +ip-26-0-175-132:679635:680078 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679635:680078 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679637:680077 [5] NCCL INFO Connected all trees +ip-26-0-175-132:679637:680077 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679637:680077 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132724:1133161 [1] NCCL INFO comm 0xa2f60e0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeebe2b6fcd69f481 - Init COMPLETE +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679635:680078 [3] NCCL INFO comm 0x9110780 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4df034fc67a447f8 - Init COMPLETE +ip-26-0-175-132:679636:680074 [4] NCCL INFO Connected all trees +ip-26-0-175-132:679636:680074 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679636:680074 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132725:1133160 [2] NCCL INFO comm 0x91fe5e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xeebe2b6fcd69f481 - Init COMPLETE +ip-26-0-175-132:679637:680077 [5] NCCL INFO comm 0xa4de580 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4df034fc67a447f8 - Init COMPLETE +ip-26-0-166-214:1132729:1133158 [6] NCCL INFO comm 0x9314aa0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xeebe2b6fcd69f481 - Init COMPLETE +ip-26-0-175-132:679636:680074 [4] NCCL INFO comm 0x9406140 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4df034fc67a447f8 - Init COMPLETE +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Connected all rings +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Connected all rings +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Connected all rings +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Connected all rings +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Connected all rings +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Connected all rings +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133162 [3] NCCL INFO comm 0x9ac3260 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xeebe2b6fcd69f481 - Init COMPLETE +ip-26-0-166-214:1132728:1133159 [5] NCCL INFO comm 0xa0df9f0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeebe2b6fcd69f481 - Init COMPLETE +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Connected all rings +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Connected all rings +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132727:1133164 [4] NCCL INFO comm 0x998b680 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xeebe2b6fcd69f481 - Init COMPLETE +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO Connected all trees +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632408:1632832 [7] NCCL INFO comm 0x8ed8540 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x25a5be1d9cb5a76f - Init COMPLETE +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO Connected all trees +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624156:3624585 [7] NCCL INFO comm 0x8aa94a0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x44bc682292fe124c - Init COMPLETE +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO Connected all trees +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632401:1632829 [0] NCCL INFO comm 0xa42e950 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x25a5be1d9cb5a76f - Init COMPLETE +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO Connected all trees +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO Connected all trees +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632402:1632833 [1] NCCL INFO comm 0x95317c0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x25a5be1d9cb5a76f - Init COMPLETE +ip-26-0-174-196:3624149:3624578 [0] NCCL INFO comm 0x8d45540 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x44bc682292fe124c - Init COMPLETE +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO Connected all trees +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO Connected all trees +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO Connected all trees +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624150:3624581 [1] NCCL INFO comm 0x91f4e00 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x44bc682292fe124c - Init COMPLETE +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO Connected all trees +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632407:1632830 [6] NCCL INFO comm 0x976c220 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x25a5be1d9cb5a76f - Init COMPLETE +ip-26-0-174-196:3624151:3624584 [2] NCCL INFO comm 0x8b35540 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44bc682292fe124c - Init COMPLETE +ip-26-0-174-196:3624155:3624583 [6] NCCL INFO comm 0xa616280 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x44bc682292fe124c - Init COMPLETE +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO Connected all trees +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO Connected all trees +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO Connected all trees +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632404:1632835 [3] NCCL INFO comm 0x9a9eac0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x25a5be1d9cb5a76f - Init COMPLETE +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO Connected all trees +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO Connected all trees +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624152:3624579 [3] NCCL INFO comm 0x939afc0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44bc682292fe124c - Init COMPLETE +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO Connected all trees +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632403:1632834 [2] NCCL INFO comm 0x86e4cd0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x25a5be1d9cb5a76f - Init COMPLETE +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO Connected all trees +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624154:3624582 [5] NCCL INFO comm 0x8b95980 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x44bc682292fe124c - Init COMPLETE +ip-26-0-168-95:1632406:1632831 [5] NCCL INFO comm 0x8ca55c0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x25a5be1d9cb5a76f - Init COMPLETE +ip-26-0-168-95:1632405:1632836 [4] NCCL INFO comm 0x8bb3a20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x25a5be1d9cb5a76f - Init COMPLETE +ip-26-0-174-196:3624153:3624580 [4] NCCL INFO comm 0x9196880 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x44bc682292fe124c - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Using network Libfabric +ip-26-0-175-132:679632:680116 [0] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Using network Libfabric +ip-26-0-175-132:679634:680117 [2] NCCL INFO Using network Libfabric +ip-26-0-175-132:679638:680112 [6] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Using network Libfabric +ip-26-0-175-132:679636:680115 [4] NCCL INFO Using network Libfabric +ip-26-0-175-132:679635:680120 [3] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Using network Libfabric +ip-26-0-175-132:679639:680118 [7] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Using network Libfabric +ip-26-0-175-132:679637:680121 [5] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Using network Libfabric +ip-26-0-175-132:679633:680119 [1] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Using network Libfabric +ip-26-0-175-132:679635:680120 [3] NCCL INFO comm 0xa4135e0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x47ee9543f860c4e1 - Init START +ip-26-0-175-132:679634:680117 [2] NCCL INFO comm 0x9e1c3e0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x47ee9543f860c4e1 - Init START +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO comm 0xaa70400 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x72799b18d85c46c4 - Init START +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO comm 0xa1db070 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x72799b18d85c46c4 - Init START +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Using network Libfabric +ip-26-0-175-132:679638:680112 [6] NCCL INFO comm 0xc216b80 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9453e30c017d770e - Init START +ip-26-0-175-132:679639:680118 [7] NCCL INFO comm 0xb0a3040 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9453e30c017d770e - Init START +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO comm 0xa69c4b0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5c80e38fb7c1cdc4 - Init START +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO comm 0x9e38f00 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5c80e38fb7c1cdc4 - Init START +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Using network Libfabric +ip-26-0-175-132:679634:680117 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679635:680120 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679636:680115 [4] NCCL INFO comm 0xb149e20 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2ba4ba6e59bb1500 - Init START +ip-26-0-175-132:679637:680121 [5] NCCL INFO comm 0xc220bb0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2ba4ba6e59bb1500 - Init START +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Using network Libfabric +ip-26-0-175-132:679632:680116 [0] NCCL INFO comm 0xa8787f0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4eb362f63847e38c - Init START +ip-26-0-175-132:679633:680119 [1] NCCL INFO comm 0xa526c00 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4eb362f63847e38c - Init START +ip-26-0-175-132:679639:680118 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679638:680112 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO comm 0xb730fe0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xba52df51466d2be9 - Init START +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO comm 0xb278a90 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xba52df51466d2be9 - Init START +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO comm 0xa048ca0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x41e46b8b6da38b9a - Init START +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO comm 0xa4f7b30 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x41e46b8b6da38b9a - Init START +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO comm 0x9e90e60 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcc9c0dfe5b880671 - Init START +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO comm 0xa492ec0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcc9c0dfe5b880671 - Init START +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679636:680115 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO comm 0xb90f9b0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x23525be02348e02e - Init START +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO comm 0x9da49e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x23525be02348e02e - Init START +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Using network Libfabric +ip-26-0-175-132:679637:680121 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Using network Libfabric +ip-26-0-175-132:679633:680119 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Using network Libfabric +ip-26-0-175-132:679632:680116 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO comm 0x9fa7ae0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcf41b52fc8d03f5a - Init START +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO comm 0x9eb4080 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcf41b52fc8d03f5a - Init START +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO comm 0xa7412a0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x338bcef92412c138 - Init START +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO comm 0xb7bbac0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x338bcef92412c138 - Init START +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO comm 0xaf8a6e0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x90e437a923071260 - Init START +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO comm 0xa6c5d70 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x90e437a923071260 - Init START +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO comm 0xa619f20 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbea2cf7fb69a47d8 - Init START +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO comm 0xa5dc800 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbea2cf7fb69a47d8 - Init START +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO comm 0xb3e1c40 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2a6233e603b9f182 - Init START +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO comm 0xac8c450 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2a6233e603b9f182 - Init START +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO comm 0x99e7b80 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9454dba04098456d - Init START +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO comm 0xada14c0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9454dba04098456d - Init START +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679634:680117 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-175-132:679632:680116 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-175-132:679639:680118 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-175-132:679635:680120 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-175-132:679635:680120 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-175-132:679635:680120 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-175-132:679634:680117 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-175-132:679634:680117 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679633:680119 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-175-132:679633:680119 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-175-132:679633:680119 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679632:680116 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-175-132:679632:680116 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-175-132:679636:680115 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-175-132:679638:680112 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-175-132:679637:680121 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679639:680118 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-175-132:679639:680118 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-175-132:679638:680112 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-175-132:679638:680112 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-175-132:679636:680115 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-175-132:679637:680121 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-175-132:679636:680115 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679637:680121 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680117 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679635:680120 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680120 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680120 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680120 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680120 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680120 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680120 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680120 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680116 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679633:680119 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680119 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680119 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680119 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680119 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680119 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680119 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680119 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679639:680118 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680118 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680118 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680118 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680118 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680118 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679639:680118 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679639:680118 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680121 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679637:680121 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679636:680115 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679637:680121 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679637:680121 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679637:680121 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679638:680112 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679637:680121 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680121 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680121 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679634:680117 [2] NCCL INFO Connected all rings +ip-26-0-175-132:679634:680117 [2] NCCL INFO Connected all trees +ip-26-0-175-132:679635:680120 [3] NCCL INFO Connected all rings +ip-26-0-175-132:679635:680120 [3] NCCL INFO Connected all trees +ip-26-0-175-132:679635:680120 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679635:680120 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679634:680117 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679634:680117 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Connected all rings +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO Connected all trees +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Connected all rings +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO Connected all trees +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679635:680120 [3] NCCL INFO comm 0xa4135e0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x47ee9543f860c4e1 - Init COMPLETE +ip-26-0-175-132:679634:680117 [2] NCCL INFO comm 0x9e1c3e0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x47ee9543f860c4e1 - Init COMPLETE +ip-26-0-174-196:3624156:3624626 [7] NCCL INFO comm 0x9da49e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x23525be02348e02e - Init COMPLETE +ip-26-0-174-196:3624155:3624621 [6] NCCL INFO comm 0xb90f9b0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x23525be02348e02e - Init COMPLETE +ip-26-0-175-132:679632:680116 [0] NCCL INFO Connected all rings +ip-26-0-175-132:679632:680116 [0] NCCL INFO Connected all trees +ip-26-0-175-132:679632:680116 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679632:680116 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679633:680119 [1] NCCL INFO Connected all rings +ip-26-0-175-132:679633:680119 [1] NCCL INFO Connected all trees +ip-26-0-175-132:679633:680119 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679633:680119 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679632:680116 [0] NCCL INFO comm 0xa8787f0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4eb362f63847e38c - Init COMPLETE +ip-26-0-175-132:679633:680119 [1] NCCL INFO comm 0xa526c00 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4eb362f63847e38c - Init COMPLETE +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Connected all rings +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO Connected all trees +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Connected all rings +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO Connected all trees +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679639:680118 [7] NCCL INFO Connected all rings +ip-26-0-175-132:679639:680118 [7] NCCL INFO Connected all trees +ip-26-0-175-132:679639:680118 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679639:680118 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679637:680121 [5] NCCL INFO Connected all rings +ip-26-0-175-132:679637:680121 [5] NCCL INFO Connected all trees +ip-26-0-175-132:679637:680121 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679637:680121 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679636:680115 [4] NCCL INFO Connected all rings +ip-26-0-175-132:679636:680115 [4] NCCL INFO Connected all trees +ip-26-0-175-132:679636:680115 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679636:680115 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679638:680112 [6] NCCL INFO Connected all rings +ip-26-0-175-132:679638:680112 [6] NCCL INFO Connected all trees +ip-26-0-175-132:679638:680112 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679638:680112 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624150:3624625 [1] NCCL INFO comm 0xa4f7b30 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x41e46b8b6da38b9a - Init COMPLETE +ip-26-0-174-196:3624149:3624624 [0] NCCL INFO comm 0xa048ca0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x41e46b8b6da38b9a - Init COMPLETE +ip-26-0-175-132:679639:680118 [7] NCCL INFO comm 0xb0a3040 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9453e30c017d770e - Init COMPLETE +ip-26-0-175-132:679637:680121 [5] NCCL INFO comm 0xc220bb0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2ba4ba6e59bb1500 - Init COMPLETE +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Connected all rings +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO Connected all trees +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Connected all rings +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO Connected all trees +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679636:680115 [4] NCCL INFO comm 0xb149e20 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2ba4ba6e59bb1500 - Init COMPLETE +ip-26-0-175-132:679638:680112 [6] NCCL INFO comm 0xc216b80 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9453e30c017d770e - Init COMPLETE +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Connected all rings +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO Connected all trees +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Connected all rings +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO Connected all trees +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Connected all rings +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO Connected all trees +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624151:3624617 [2] NCCL INFO comm 0x9e38f00 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5c80e38fb7c1cdc4 - Init COMPLETE +ip-26-0-174-196:3624152:3624623 [3] NCCL INFO comm 0xa69c4b0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5c80e38fb7c1cdc4 - Init COMPLETE +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Connected all rings +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO Connected all trees +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624153:3624618 [4] NCCL INFO comm 0xa492ec0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcc9c0dfe5b880671 - Init COMPLETE +ip-26-0-174-196:3624154:3624622 [5] NCCL INFO comm 0x9e90e60 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcc9c0dfe5b880671 - Init COMPLETE +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Connected all rings +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO Connected all trees +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Connected all rings +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO Connected all trees +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632408:1632870 [7] NCCL INFO comm 0xa1db070 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x72799b18d85c46c4 - Init COMPLETE +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Connected all rings +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO Connected all trees +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632402:1632877 [1] NCCL INFO comm 0xb278a90 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xba52df51466d2be9 - Init COMPLETE +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Connected all rings +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO Connected all trees +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Connected all rings +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO Connected all trees +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Connected all rings +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO Connected all trees +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632407:1632868 [6] NCCL INFO comm 0xaa70400 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x72799b18d85c46c4 - Init COMPLETE +ip-26-0-168-95:1632401:1632875 [0] NCCL INFO comm 0xb730fe0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xba52df51466d2be9 - Init COMPLETE +ip-26-0-168-95:1632403:1632876 [2] NCCL INFO comm 0x99e7b80 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9454dba04098456d - Init COMPLETE +ip-26-0-168-95:1632405:1632871 [4] NCCL INFO comm 0x9eb4080 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcf41b52fc8d03f5a - Init COMPLETE +ip-26-0-168-95:1632404:1632878 [3] NCCL INFO comm 0xada14c0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9454dba04098456d - Init COMPLETE +ip-26-0-168-95:1632406:1632873 [5] NCCL INFO comm 0x9fa7ae0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcf41b52fc8d03f5a - Init COMPLETE +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132728:1133204 [5] NCCL INFO comm 0xb3e1c40 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2a6233e603b9f182 - Init COMPLETE +ip-26-0-166-214:1132727:1133198 [4] NCCL INFO comm 0xac8c450 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2a6233e603b9f182 - Init COMPLETE +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132724:1133202 [1] NCCL INFO comm 0xb7bbac0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x338bcef92412c138 - Init COMPLETE +ip-26-0-166-214:1132723:1133195 [0] NCCL INFO comm 0xa7412a0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x338bcef92412c138 - Init COMPLETE +ip-26-0-166-214:1132730:1133201 [7] NCCL INFO comm 0xa5dc800 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbea2cf7fb69a47d8 - Init COMPLETE +ip-26-0-166-214:1132726:1133203 [3] NCCL INFO comm 0xaf8a6e0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x90e437a923071260 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-214:1132725:1133200 [2] NCCL INFO comm 0xa6c5d70 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x90e437a923071260 - Init COMPLETE +ip-26-0-166-214:1132729:1133199 [6] NCCL INFO comm 0xa619f20 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbea2cf7fb69a47d8 - Init COMPLETE +ip-26-0-175-132:679637:680142 [5] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Using network Libfabric +ip-26-0-175-132:679633:680141 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Using network Libfabric +ip-26-0-175-132:679632:680144 [0] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Using network Libfabric +ip-26-0-175-132:679636:680143 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO comm 0xbe25dd0 rank 1 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf551a6a2b028e85a - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-175-132:679637:680142 [5] NCCL INFO comm 0xc23bf90 rank 7 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf551a6a2b028e85a - Init START +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO comm 0xa9ec6c0 rank 3 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf551a6a2b028e85a - Init START +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO comm 0xc2001a0 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf551a6a2b028e85a - Init START +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO comm 0xaf3c390 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf551a6a2b028e85a - Init START +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO comm 0xa8d6620 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf551a6a2b028e85a - Init START +ip-26-0-175-132:679633:680141 [1] NCCL INFO comm 0xaf6b020 rank 6 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf551a6a2b028e85a - Init START +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO comm 0xb2921d0 rank 2 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf551a6a2b028e85a - Init START +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679637:680142 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679633:680141 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679635:680146 [3] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Using network Libfabric +ip-26-0-175-132:679639:680145 [7] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Using network Libfabric +ip-26-0-175-132:679636:680143 [4] NCCL INFO comm 0xb164f60 rank 7 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x11501c345b56b5c1 - Init START +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO comm 0xb185a00 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x11501c345b56b5c1 - Init START +ip-26-0-175-132:679632:680144 [0] NCCL INFO comm 0xb2bd050 rank 6 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x11501c345b56b5c1 - Init START +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO comm 0xb6d04d0 rank 1 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x11501c345b56b5c1 - Init START +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO comm 0xa8f8590 rank 3 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x11501c345b56b5c1 - Init START +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO comm 0xc1753b0 rank 2 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x11501c345b56b5c1 - Init START +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO comm 0xaed8a30 rank 5 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x11501c345b56b5c1 - Init START +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO comm 0xaa8d210 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x11501c345b56b5c1 - Init START +ip-26-0-175-132:679638:680148 [6] NCCL INFO Using network Libfabric +ip-26-0-175-132:679634:680147 [2] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-175-132:679636:680143 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679632:680144 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO comm 0xb10a080 rank 0 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a541a2e29c98424 - Init START +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO comm 0xb05e3f0 rank 1 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1a541a2e29c98424 - Init START +ip-26-0-175-132:679638:680148 [6] NCCL INFO comm 0xc231f70 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1a541a2e29c98424 - Init START +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO comm 0xb4b4c90 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1a541a2e29c98424 - Init START +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO comm 0xc355480 rank 5 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1a541a2e29c98424 - Init START +ip-26-0-175-132:679634:680147 [2] NCCL INFO comm 0xa860be0 rank 6 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a541a2e29c98424 - Init START +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO comm 0xa42bcb0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a541a2e29c98424 - Init START +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO comm 0xa87d2d0 rank 4 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a541a2e29c98424 - Init START +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679634:680147 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679638:680148 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679639:680145 [7] NCCL INFO comm 0xb0bdeb0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7180e6a6246f552 - Init START +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO comm 0xb9cec40 rank 0 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7180e6a6246f552 - Init START +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO comm 0xb020b30 rank 1 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7180e6a6246f552 - Init START +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679635:680146 [3] NCCL INFO comm 0xae57960 rank 6 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7180e6a6246f552 - Init START +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO comm 0xac1f160 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7180e6a6246f552 - Init START +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO comm 0xb7e56f0 rank 2 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7180e6a6246f552 - Init START +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO comm 0xb0e0d00 rank 4 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7180e6a6246f552 - Init START +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO comm 0xa7ea090 rank 5 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7180e6a6246f552 - Init START +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679639:680145 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679635:680146 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-175-132:679632:680144 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-175-132:679639:680145 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-175-132:679638:680148 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-175-132:679634:680147 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-175-132:679637:680142 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-175-132:679635:680146 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-175-132:679633:680141 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-175-132:679636:680143 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-175-132:679637:680142 [5] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-175-132:679637:680142 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-175-132:679633:680141 [1] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-175-132:679633:680141 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679638:680148 [6] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-175-132:679638:680148 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679634:680147 [2] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-175-132:679634:680147 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679636:680143 [4] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-175-132:679636:680143 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679632:680144 [0] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-175-132:679632:680144 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679635:680146 [3] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-175-132:679635:680146 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679639:680145 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679639:680145 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 00/0 : 7[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 01/0 : 7[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 02/0 : 7[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 03/0 : 7[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 00/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 01/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 02/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 03/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 00/0 : 3[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 01/0 : 3[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 02/0 : 3[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 03/0 : 3[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 00/0 : 4[1] -> 5[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 01/0 : 4[1] -> 5[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 02/0 : 4[1] -> 5[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 03/0 : 4[1] -> 5[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 00/0 : 6[1] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 01/0 : 6[1] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 00/0 : 5[5] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 02/0 : 6[1] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 01/0 : 5[5] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 03/0 : 6[1] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 02/0 : 5[5] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 00/0 : 7[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 03/0 : 5[5] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 01/0 : 7[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 00/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 00/0 : 6[1] -> 7[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 02/0 : 7[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 01/0 : 6[1] -> 7[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 03/0 : 7[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 01/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 02/0 : 6[1] -> 7[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 02/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 03/0 : 6[1] -> 7[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 03/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 00/0 : 3[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 01/0 : 3[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 02/0 : 3[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 03/0 : 3[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 00/0 : 4[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 01/0 : 4[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 02/0 : 4[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 03/0 : 4[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 00/0 : 5[6] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 01/0 : 5[6] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 02/0 : 5[6] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 03/0 : 5[6] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 00/0 : 6[2] -> 7[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 01/0 : 6[2] -> 7[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 02/0 : 6[2] -> 7[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 03/0 : 6[2] -> 7[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 00/0 : 7[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 01/0 : 7[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 02/0 : 7[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 03/0 : 7[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 00/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 01/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 02/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 03/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 00/0 : 4[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 01/0 : 4[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 02/0 : 4[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 03/0 : 4[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 00/0 : 5[6] -> 6[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 01/0 : 5[6] -> 6[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 02/0 : 5[6] -> 6[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 03/0 : 5[6] -> 6[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 00/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 01/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 02/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 03/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 00/0 : 6[2] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 00/0 : 3[6] -> 4[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 01/0 : 3[6] -> 4[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 01/0 : 6[2] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 02/0 : 3[6] -> 4[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 02/0 : 6[2] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 03/0 : 3[6] -> 4[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 03/0 : 6[2] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 00/0 : 7[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 01/0 : 7[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 02/0 : 7[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 03/0 : 7[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 00/0 : 3[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 01/0 : 3[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 02/0 : 3[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 03/0 : 3[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 00/0 : 4[2] -> 5[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 01/0 : 4[2] -> 5[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 02/0 : 4[2] -> 5[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 03/0 : 4[2] -> 5[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 00/0 : 7[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 01/0 : 7[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 02/0 : 7[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 03/0 : 7[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 00/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 01/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 02/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 03/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 00/0 : 3[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 01/0 : 3[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 02/0 : 3[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 03/0 : 3[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 00/0 : 6[0] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 01/0 : 6[0] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 02/0 : 6[0] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 03/0 : 6[0] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 00/0 : 7[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 01/0 : 7[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 02/0 : 7[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 00/0 : 4[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 03/0 : 7[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 01/0 : 4[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 02/0 : 4[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 00/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 00/0 : 5[4] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 03/0 : 4[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 01/0 : 5[4] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 01/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 00/0 : 5[4] -> 6[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 02/0 : 5[4] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 02/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 01/0 : 5[4] -> 6[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 03/0 : 5[4] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 03/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 02/0 : 5[4] -> 6[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 00/0 : 6[0] -> 7[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 00/0 : 3[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 03/0 : 5[4] -> 6[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 01/0 : 6[0] -> 7[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 01/0 : 3[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 02/0 : 6[0] -> 7[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 03/0 : 6[0] -> 7[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 02/0 : 3[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 03/0 : 3[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 00/0 : 4[0] -> 5[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 01/0 : 4[0] -> 5[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 02/0 : 4[0] -> 5[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 03/0 : 4[0] -> 5[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 00/0 : 3[7] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 01/0 : 3[7] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 02/0 : 3[7] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 03/0 : 3[7] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 00/0 : 4[3] -> 5[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 01/0 : 4[3] -> 5[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 00/0 : 7[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 02/0 : 4[3] -> 5[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 01/0 : 7[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 03/0 : 4[3] -> 5[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 02/0 : 7[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 03/0 : 7[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 00/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 01/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 02/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 03/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 00/0 : 3[7] -> 4[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 01/0 : 3[7] -> 4[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 02/0 : 3[7] -> 4[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 03/0 : 3[7] -> 4[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 00/0 : 6[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 01/0 : 6[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 02/0 : 6[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 03/0 : 6[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 00/0 : 4[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 01/0 : 4[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 02/0 : 4[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 03/0 : 4[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 00/0 : 5[7] -> 6[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 01/0 : 5[7] -> 6[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 02/0 : 5[7] -> 6[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 03/0 : 5[7] -> 6[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 00/0 : 5[7] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 01/0 : 5[7] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 02/0 : 5[7] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 03/0 : 5[7] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 00/0 : 6[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 01/0 : 6[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 02/0 : 6[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 03/0 : 6[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 00/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 01/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 02/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 03/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Connected all rings +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 00/0 : 2[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 01/0 : 2[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 00/0 : 4[1] -> 6[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 01/0 : 4[1] -> 6[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Connected all rings +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 02/0 : 3[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 03/0 : 3[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Connected all rings +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 00/0 : 2[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 01/0 : 2[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 00/0 : 4[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 01/0 : 4[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Connected all rings +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 02/0 : 3[5] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 03/0 : 3[5] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Connected all rings +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 02/0 : 7[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 02/0 : 3[6] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 03/0 : 7[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 03/0 : 3[6] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Connected all rings +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 00/0 : 4[1] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 01/0 : 4[1] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 00/0 : 6[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 01/0 : 6[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 00/0 : 6[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 01/0 : 6[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 00/0 : 4[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 01/0 : 4[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 00/0 : 7[5] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 01/0 : 7[5] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 02/0 : 5[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 00/0 : 6[1] -> 5[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Channel 03/0 : 5[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 00/0 : 3[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 01/0 : 6[1] -> 5[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 01/0 : 3[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 02/0 : 6[1] -> 5[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680141 [1] NCCL INFO Channel 03/0 : 6[1] -> 5[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Connected all rings +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Connected all rings +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 00/0 : 4[2] -> 6[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 01/0 : 4[2] -> 6[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Connected all rings +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 00/0 : 4[2] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 01/0 : 4[2] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 00/0 : 6[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 01/0 : 6[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Connected all rings +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 02/0 : 3[4] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 03/0 : 3[4] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Connected all rings +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 00/0 : 4[0] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 01/0 : 4[0] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Connected all rings +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 00/0 : 2[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 01/0 : 2[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 00/0 : 4[0] -> 6[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 01/0 : 4[0] -> 6[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 00/0 : 6[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 01/0 : 6[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Connected all rings +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 00/0 : 2[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 01/0 : 2[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 00/0 : 4[3] -> 6[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 01/0 : 4[3] -> 6[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Connected all rings +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 00/0 : 4[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 01/0 : 4[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 00/0 : 0[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 01/0 : 0[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Connected all rings +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 02/0 : 3[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 03/0 : 3[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 02/0 : 7[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 03/0 : 7[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 02/0 : 3[5] -> 7[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 03/0 : 3[5] -> 7[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 00/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 01/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 02/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 03/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 00/0 : 6[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 01/0 : 6[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Connected all rings +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 02/0 : 6[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 02/0 : 3[6] -> 5[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 03/0 : 6[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 02/0 : 3[4] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 03/0 : 3[6] -> 5[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 03/0 : 3[4] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Connected all rings +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 00/0 : 7[5] -> 6[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 02/0 : 7[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Channel 01/0 : 7[5] -> 6[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 03/0 : 7[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 02/0 : 5[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 03/0 : 5[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 00/0 : 4[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 01/0 : 4[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 02/0 : 5[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 00/0 : 7[6] -> 6[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 03/0 : 5[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680148 [6] NCCL INFO Channel 01/0 : 7[6] -> 6[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 00/0 : 0[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Connected all rings +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 01/0 : 0[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 00/0 : 4[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 01/0 : 4[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 00/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 00/0 : 6[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 01/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Connected all rings +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 01/0 : 6[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 02/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 00/0 : 4[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 03/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 02/0 : 3[7] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 01/0 : 4[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 03/0 : 3[7] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 02/0 : 5[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 00/0 : 3[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 00/0 : 6[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Channel 03/0 : 5[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 01/0 : 6[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Connected all rings +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 01/0 : 3[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 02/0 : 6[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 03/0 : 6[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 00/0 : 4[3] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 02/0 : 5[6] -> 4[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 01/0 : 4[3] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Channel 03/0 : 5[6] -> 4[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 00/0 : 6[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 01/0 : 6[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 00/0 : 7[6] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 01/0 : 7[6] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 00/0 : 6[2] -> 5[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 01/0 : 6[2] -> 5[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 02/0 : 6[2] -> 5[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680147 [2] NCCL INFO Channel 03/0 : 6[2] -> 5[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 00/0 : 4[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 01/0 : 4[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 00/0 : 0[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 01/0 : 0[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Connected all rings +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 00/0 : 2[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 01/0 : 2[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 00/0 : 4[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 01/0 : 4[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 00/0 : 6[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 01/0 : 6[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 00/0 : 4[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 01/0 : 4[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 00/0 : 7[4] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 01/0 : 7[4] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 02/0 : 5[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 00/0 : 3[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 00/0 : 6[0] -> 5[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Channel 03/0 : 5[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 01/0 : 3[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 01/0 : 6[0] -> 5[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 02/0 : 6[0] -> 5[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680144 [0] NCCL INFO Channel 03/0 : 6[0] -> 5[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Connected all rings +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 00/0 : 2[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 01/0 : 2[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 00/0 : 4[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 01/0 : 4[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 00/0 : 0[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 01/0 : 0[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 00/0 : 4[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 01/0 : 4[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 00/0 : 6[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 01/0 : 6[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 00/0 : 4[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 01/0 : 4[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 00/0 : 7[7] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 02/0 : 5[7] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 01/0 : 7[7] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 00/0 : 3[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Channel 03/0 : 5[7] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 00/0 : 6[3] -> 5[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 01/0 : 3[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 01/0 : 6[3] -> 5[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 02/0 : 6[3] -> 5[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680146 [3] NCCL INFO Channel 03/0 : 6[3] -> 5[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Connected all rings +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 02/0 : 3[4] -> 5[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 03/0 : 3[4] -> 5[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 02/0 : 7[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 03/0 : 7[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 02/0 : 3[4] -> 7[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 03/0 : 3[4] -> 7[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 02/0 : 5[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 03/0 : 5[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 02/0 : 5[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 03/0 : 5[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 00/0 : 6[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 00/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 01/0 : 6[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 01/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 02/0 : 6[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 02/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 03/0 : 6[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 03/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 02/0 : 5[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Channel 03/0 : 5[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 00/0 : 7[4] -> 6[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680143 [4] NCCL INFO Channel 01/0 : 7[4] -> 6[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Connected all rings +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 02/0 : 3[7] -> 5[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 03/0 : 3[7] -> 5[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 02/0 : 5[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 03/0 : 5[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 02/0 : 5[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 03/0 : 5[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680145 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 00/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 01/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 02/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 03/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 00/0 : 6[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 01/0 : 6[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 02/0 : 6[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 03/0 : 6[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 02/0 : 5[7] -> 4[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Channel 03/0 : 5[7] -> 4[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679637:680142 [5] NCCL INFO Connected all trees +ip-26-0-175-132:679637:680142 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679637:680142 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679637:680142 [5] NCCL INFO comm 0xc23bf90 rank 7 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf551a6a2b028e85a - Init COMPLETE +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679638:680148 [6] NCCL INFO Connected all trees +ip-26-0-175-132:679638:680148 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679638:680148 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132724:1133226 [1] NCCL INFO comm 0xc2001a0 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf551a6a2b028e85a - Init COMPLETE +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679639:680145 [7] NCCL INFO Connected all trees +ip-26-0-175-132:679639:680145 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679639:680145 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679638:680148 [6] NCCL INFO comm 0xc231f70 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1a541a2e29c98424 - Init COMPLETE +ip-26-0-166-214:1132725:1133234 [2] NCCL INFO comm 0xb10a080 rank 0 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a541a2e29c98424 - Init COMPLETE +ip-26-0-175-132:679639:680145 [7] NCCL INFO comm 0xb0bdeb0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7180e6a6246f552 - Init COMPLETE +ip-26-0-175-132:679636:680143 [4] NCCL INFO Connected all trees +ip-26-0-175-132:679636:680143 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679636:680143 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679633:680141 [1] NCCL INFO Connected all trees +ip-26-0-175-132:679633:680141 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679633:680141 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679636:680143 [4] NCCL INFO comm 0xb164f60 rank 7 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x11501c345b56b5c1 - Init COMPLETE +ip-26-0-166-214:1132726:1133231 [3] NCCL INFO comm 0xb9cec40 rank 0 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7180e6a6246f552 - Init COMPLETE +ip-26-0-175-132:679633:680141 [1] NCCL INFO comm 0xaf6b020 rank 6 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf551a6a2b028e85a - Init COMPLETE +ip-26-0-175-132:679634:680147 [2] NCCL INFO Connected all trees +ip-26-0-175-132:679634:680147 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679634:680147 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO Connected all trees +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO Connected all trees +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO Connected all trees +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679634:680147 [2] NCCL INFO comm 0xa860be0 rank 6 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a541a2e29c98424 - Init COMPLETE +ip-26-0-174-196:3624154:3624646 [5] NCCL INFO comm 0xa8d6620 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf551a6a2b028e85a - Init COMPLETE +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO Connected all trees +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679632:680144 [0] NCCL INFO Connected all trees +ip-26-0-175-132:679632:680144 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679632:680144 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132729:1133235 [6] NCCL INFO comm 0xb05e3f0 rank 1 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1a541a2e29c98424 - Init COMPLETE +ip-26-0-166-214:1132728:1133225 [5] NCCL INFO comm 0xbe25dd0 rank 1 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf551a6a2b028e85a - Init COMPLETE +ip-26-0-168-95:1632402:1632898 [1] NCCL INFO comm 0xb2921d0 rank 2 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf551a6a2b028e85a - Init COMPLETE +ip-26-0-174-196:3624155:3624652 [6] NCCL INFO comm 0xc355480 rank 5 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1a541a2e29c98424 - Init COMPLETE +ip-26-0-168-95:1632403:1632904 [2] NCCL INFO comm 0xa42bcb0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a541a2e29c98424 - Init COMPLETE +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679632:680144 [0] NCCL INFO comm 0xb2bd050 rank 6 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x11501c345b56b5c1 - Init COMPLETE +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132723:1133228 [0] NCCL INFO comm 0xb185a00 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x11501c345b56b5c1 - Init COMPLETE +ip-26-0-166-214:1132727:1133229 [4] NCCL INFO comm 0xb6d04d0 rank 1 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x11501c345b56b5c1 - Init COMPLETE +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO Connected all trees +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624156:3624650 [7] NCCL INFO comm 0xa7ea090 rank 5 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7180e6a6246f552 - Init COMPLETE +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO Connected all trees +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132730:1133232 [7] NCCL INFO comm 0xb020b30 rank 1 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7180e6a6246f552 - Init COMPLETE +ip-26-0-168-95:1632404:1632901 [3] NCCL INFO comm 0xb7e56f0 rank 2 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7180e6a6246f552 - Init COMPLETE +ip-26-0-175-132:679635:680146 [3] NCCL INFO Connected all trees +ip-26-0-175-132:679635:680146 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-175-132:679635:680146 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO Connected all trees +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679635:680146 [3] NCCL INFO comm 0xae57960 rank 6 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7180e6a6246f552 - Init COMPLETE +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO Connected all trees +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO Connected all trees +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632401:1632900 [0] NCCL INFO comm 0xc1753b0 rank 2 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x11501c345b56b5c1 - Init COMPLETE +ip-26-0-174-196:3624150:3624645 [1] NCCL INFO comm 0xaf3c390 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf551a6a2b028e85a - Init COMPLETE +ip-26-0-174-196:3624151:3624651 [2] NCCL INFO comm 0xa87d2d0 rank 4 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a541a2e29c98424 - Init COMPLETE +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO Connected all trees +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO Connected all trees +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO Connected all trees +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO Connected all trees +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624153:3624648 [4] NCCL INFO comm 0xaed8a30 rank 5 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x11501c345b56b5c1 - Init COMPLETE +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO Connected all trees +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624149:3624647 [0] NCCL INFO comm 0xaa8d210 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x11501c345b56b5c1 - Init COMPLETE +ip-26-0-168-95:1632407:1632903 [6] NCCL INFO comm 0xb4b4c90 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1a541a2e29c98424 - Init COMPLETE +ip-26-0-168-95:1632406:1632897 [5] NCCL INFO comm 0xa9ec6c0 rank 3 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf551a6a2b028e85a - Init COMPLETE +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO Connected all trees +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624152:3624649 [3] NCCL INFO comm 0xb0e0d00 rank 4 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7180e6a6246f552 - Init COMPLETE +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO Connected all trees +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632408:1632902 [7] NCCL INFO comm 0xac1f160 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7180e6a6246f552 - Init COMPLETE +ip-26-0-168-95:1632405:1632899 [4] NCCL INFO comm 0xa8f8590 rank 3 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x11501c345b56b5c1 - Init COMPLETE +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO comm 0xc22aff0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x846fa1e21b593098 - Init START +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO comm 0xb9fa2f0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x846fa1e21b593098 - Init START +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO comm 0xb134f60 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x289aae1cbd4b30e2 - Init START +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO comm 0xb1b4640 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x289aae1cbd4b30e2 - Init START +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Using network Libfabric +ip-26-0-175-132:679633:680169 [1] NCCL INFO Using network Libfabric +ip-26-0-175-132:679635:680170 [3] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO comm 0xaf746e0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5e64364a724cd64c - Init START +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO comm 0xb119680 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5e64364a724cd64c - Init START +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679633:680169 [1] NCCL INFO comm 0xafa3e50 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x70d69e63510eda13 - Init START +ip-26-0-175-132:679635:680170 [3] NCCL INFO comm 0xae901a0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x70d69e63510eda13 - Init START +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679633:680169 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679635:680170 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO comm 0xb2b12e0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbf5800bdf6acaf22 - Init START +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO comm 0xb81d460 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbf5800bdf6acaf22 - Init START +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO comm 0xaa25050 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x100c80acc2d76763 - Init START +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO comm 0xac57720 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x100c80acc2d76763 - Init START +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO comm 0xa823650 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd23bb6dfd2102ecf - Init START +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO comm 0xa910410 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd23bb6dfd2102ecf - Init START +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679637:680172 [5] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO comm 0xbe5d6f0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x734a44e77b01be6a - Init START +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO comm 0xb0589d0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x734a44e77b01be6a - Init START +ip-26-0-175-132:679639:680173 [7] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Using network Libfabric +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO comm 0xb4ed600 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb8369d95b186c37d - Init START +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO comm 0xa931190 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb8369d95b186c37d - Init START +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO comm 0xc1ad6e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9720c869b6f5bb29 - Init START +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO comm 0xa4655c0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9720c869b6f5bb29 - Init START +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Using network Libfabric +ip-26-0-175-132:679637:680172 [5] NCCL INFO comm 0xc25b0a0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde16ab8b09284c19 - Init START +ip-26-0-175-132:679639:680173 [7] NCCL INFO comm 0xb0dce40 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xde16ab8b09284c19 - Init START +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679637:680172 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679639:680173 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO comm 0xaac5710 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x202ddfef122be1f9 - Init START +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO comm 0xa8b62f0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x202ddfef122be1f9 - Init START +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO comm 0xb708830 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc1c696a369a239ff - Init START +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO comm 0xb097690 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc1c696a369a239ff - Init START +ip-26-0-175-132:679636:680175 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679638:680176 [6] NCCL INFO Using network Libfabric +ip-26-0-175-132:679632:680178 [0] NCCL INFO Using network Libfabric +ip-26-0-175-132:679636:680175 [4] NCCL INFO comm 0xb184070 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6500da7d75f068a7 - Init START +ip-26-0-175-132:679638:680176 [6] NCCL INFO comm 0xc251080 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6500da7d75f068a7 - Init START +ip-26-0-175-132:679634:680179 [2] NCCL INFO Using network Libfabric +ip-26-0-175-132:679636:680175 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679638:680176 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679634:680179 [2] NCCL INFO comm 0xa899570 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd82c4df8a44fb1ee - Init START +ip-26-0-175-132:679632:680178 [0] NCCL INFO comm 0xb2f5670 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd82c4df8a44fb1ee - Init START +ip-26-0-175-132:679634:680179 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-175-132:679632:680178 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Using network Libfabric +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO comm 0xc3900a0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x536ed7e1ec201b5b - Init START +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO comm 0xaf120f0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x536ed7e1ec201b5b - Init START +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-175-132:679632:680178 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-175-132:679636:680175 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-175-132:679638:680176 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-175-132:679638:680176 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-175-132:679638:680176 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679636:680175 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-175-132:679636:680175 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679634:680179 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-175-132:679634:680179 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-175-132:679632:680178 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-175-132:679634:680179 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679632:680178 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-175-132:679635:680170 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679639:680173 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-175-132:679633:680169 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-175-132:679635:680170 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-175-132:679635:680170 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-175-132:679633:680169 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-175-132:679633:680169 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-175-132:679637:680172 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-175-132:679639:680173 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-175-132:679639:680173 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-175-132:679637:680172 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-175-132:679637:680172 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679636:680175 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-175-132:679638:680176 [6] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680176 [6] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680176 [6] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680176 [6] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680176 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680176 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680176 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679638:680176 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679634:680179 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680179 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680179 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680179 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680179 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680179 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680179 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679634:680179 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-175-132:679632:680178 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679633:680169 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-175-132:679635:680170 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680170 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680170 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680170 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680170 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680170 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680170 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-175-132:679635:680170 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-175-132:679639:680173 [7] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680173 [7] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680173 [7] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680173 [7] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680173 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680173 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680173 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679639:680173 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-175-132:679637:680172 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 00/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 01/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 02/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 03/0 : 1[7] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Connected all rings +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO Connected all trees +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624156:3624678 [7] NCCL INFO comm 0xa823650 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd23bb6dfd2102ecf - Init COMPLETE +ip-26-0-175-132:679638:680176 [6] NCCL INFO Connected all rings +ip-26-0-175-132:679638:680176 [6] NCCL INFO Connected all trees +ip-26-0-175-132:679638:680176 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679638:680176 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679636:680175 [4] NCCL INFO Connected all rings +ip-26-0-175-132:679636:680175 [4] NCCL INFO Connected all trees +ip-26-0-175-132:679636:680175 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679636:680175 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679632:680178 [0] NCCL INFO Connected all rings +ip-26-0-175-132:679632:680178 [0] NCCL INFO Connected all trees +ip-26-0-175-132:679632:680178 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679632:680178 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679634:680179 [2] NCCL INFO Connected all rings +ip-26-0-175-132:679634:680179 [2] NCCL INFO Connected all trees +ip-26-0-175-132:679634:680179 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679634:680179 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679638:680176 [6] NCCL INFO comm 0xc251080 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6500da7d75f068a7 - Init COMPLETE +ip-26-0-175-132:679636:680175 [4] NCCL INFO comm 0xb184070 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6500da7d75f068a7 - Init COMPLETE +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Connected all rings +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO Connected all trees +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679632:680178 [0] NCCL INFO comm 0xb2f5670 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd82c4df8a44fb1ee - Init COMPLETE +ip-26-0-175-132:679634:680179 [2] NCCL INFO comm 0xa899570 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd82c4df8a44fb1ee - Init COMPLETE +ip-26-0-174-196:3624154:3624677 [5] NCCL INFO comm 0xa910410 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd23bb6dfd2102ecf - Init COMPLETE +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Connected all rings +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO Connected all trees +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Connected all rings +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO Connected all trees +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624152:3624675 [3] NCCL INFO comm 0xb119680 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5e64364a724cd64c - Init COMPLETE +ip-26-0-174-196:3624150:3624674 [1] NCCL INFO comm 0xaf746e0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5e64364a724cd64c - Init COMPLETE +ip-26-0-175-132:679635:680170 [3] NCCL INFO Connected all rings +ip-26-0-175-132:679635:680170 [3] NCCL INFO Connected all trees +ip-26-0-175-132:679635:680170 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679635:680170 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679639:680173 [7] NCCL INFO Connected all rings +ip-26-0-175-132:679639:680173 [7] NCCL INFO Connected all trees +ip-26-0-175-132:679639:680173 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679639:680173 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679637:680172 [5] NCCL INFO Connected all rings +ip-26-0-175-132:679637:680172 [5] NCCL INFO Connected all trees +ip-26-0-175-132:679637:680172 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679637:680172 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Connected all rings +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO Connected all trees +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679633:680169 [1] NCCL INFO Connected all rings +ip-26-0-175-132:679633:680169 [1] NCCL INFO Connected all trees +ip-26-0-175-132:679633:680169 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-175-132:679633:680169 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Connected all rings +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO Connected all trees +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Connected all rings +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO Connected all trees +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Connected all rings +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO Connected all trees +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-175-132:679635:680170 [3] NCCL INFO comm 0xae901a0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x70d69e63510eda13 - Init COMPLETE +ip-26-0-175-132:679639:680173 [7] NCCL INFO comm 0xb0dce40 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xde16ab8b09284c19 - Init COMPLETE +ip-26-0-175-132:679637:680172 [5] NCCL INFO comm 0xc25b0a0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde16ab8b09284c19 - Init COMPLETE +ip-26-0-175-132:679633:680169 [1] NCCL INFO comm 0xafa3e50 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x70d69e63510eda13 - Init COMPLETE +ip-26-0-174-196:3624149:3624680 [0] NCCL INFO comm 0xaac5710 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x202ddfef122be1f9 - Init COMPLETE +ip-26-0-174-196:3624151:3624681 [2] NCCL INFO comm 0xa8b62f0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x202ddfef122be1f9 - Init COMPLETE +ip-26-0-174-196:3624155:3624684 [6] NCCL INFO comm 0xc3900a0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x536ed7e1ec201b5b - Init COMPLETE +ip-26-0-174-196:3624153:3624683 [4] NCCL INFO comm 0xaf120f0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x536ed7e1ec201b5b - Init COMPLETE +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Connected all rings +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO Connected all trees +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Connected all rings +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO Connected all trees +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Connected all rings +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO Connected all trees +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Connected all rings +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO Connected all trees +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632405:1632931 [4] NCCL INFO comm 0xa931190 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb8369d95b186c37d - Init COMPLETE +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Connected all rings +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO Connected all trees +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Connected all rings +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO Connected all trees +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632407:1632934 [6] NCCL INFO comm 0xb4ed600 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb8369d95b186c37d - Init COMPLETE +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Connected all rings +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO Connected all trees +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Connected all rings +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO Connected all trees +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632402:1632925 [1] NCCL INFO comm 0xb2b12e0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbf5800bdf6acaf22 - Init COMPLETE +ip-26-0-168-95:1632403:1632935 [2] NCCL INFO comm 0xa4655c0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9720c869b6f5bb29 - Init COMPLETE +ip-26-0-168-95:1632408:1632929 [7] NCCL INFO comm 0xac57720 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x100c80acc2d76763 - Init COMPLETE +ip-26-0-168-95:1632404:1632926 [3] NCCL INFO comm 0xb81d460 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbf5800bdf6acaf22 - Init COMPLETE +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632406:1632928 [5] NCCL INFO comm 0xaa25050 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x100c80acc2d76763 - Init COMPLETE +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1632401:1632933 [0] NCCL INFO comm 0xc1ad6e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9720c869b6f5bb29 - Init COMPLETE +ip-26-0-166-214:1132723:1133258 [0] NCCL INFO comm 0xb1b4640 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x289aae1cbd4b30e2 - Init COMPLETE +ip-26-0-166-214:1132725:1133259 [2] NCCL INFO comm 0xb134f60 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x289aae1cbd4b30e2 - Init COMPLETE +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132730:1133262 [7] NCCL INFO comm 0xb0589d0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x734a44e77b01be6a - Init COMPLETE +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132728:1133261 [5] NCCL INFO comm 0xbe5d6f0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x734a44e77b01be6a - Init COMPLETE +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1132724:1133255 [1] NCCL INFO comm 0xc22aff0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x846fa1e21b593098 - Init COMPLETE +ip-26-0-166-214:1132727:1133264 [4] NCCL INFO comm 0xb708830 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc1c696a369a239ff - Init COMPLETE +ip-26-0-166-214:1132729:1133265 [6] NCCL INFO comm 0xb097690 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc1c696a369a239ff - Init COMPLETE +ip-26-0-166-214:1132726:1133256 [3] NCCL INFO comm 0xb9fa2f0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x846fa1e21b593098 - Init COMPLETE +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 103.13 GiB. GPU 6 has a total capacty of 79.33 GiB of which 23.96 GiB is free. Including non-PyTorch memory, this process has 55.36 GiB memory in use. Of the allocated memory 51.56 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 103.13 GiB. GPU 7 has a total capacty of 79.33 GiB of which 23.92 GiB is free. Including non-PyTorch memory, this process has 55.39 GiB memory in use. Of the allocated memory 51.56 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 103.13 GiB. GPU 4 has a total capacty of 79.33 GiB of which 23.85 GiB is free. Including non-PyTorch memory, this process has 55.47 GiB memory in use. Of the allocated memory 51.56 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 103.13 GiB. GPU 5 has a total capacty of 79.33 GiB of which 23.96 GiB is free. Including non-PyTorch memory, this process has 55.36 GiB memory in use. Of the allocated memory 51.56 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + optimizer = ZeroDistributedOptimizer( +optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(param_groups_in_rank) + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 107.13 GiB. GPU 7 has a total capacty of 79.33 GiB of which 22.12 GiB is free. Including non-PyTorch memory, this process has 57.20 GiB memory in use. Of the allocated memory 53.56 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 107.13 GiB. GPU 4 has a total capacty of 79.33 GiB of which 21.88 GiB is free. Including non-PyTorch memory, this process has 57.43 GiB memory in use. Of the allocated memory 53.56 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 107.13 GiB. GPU 5 has a total capacty of 79.33 GiB of which 22.00 GiB is free. Including non-PyTorch memory, this process has 57.32 GiB memory in use. Of the allocated memory 53.56 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 107.13 GiB. GPU 6 has a total capacty of 79.33 GiB of which 22.00 GiB is free. Including non-PyTorch memory, this process has 57.32 GiB memory in use. Of the allocated memory 53.56 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = optimizer_builder(param_groups_in_rank) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ +result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 20.41 GiB is free. Including non-PyTorch memory, this process has 58.91 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 20.53 GiB is free. Including non-PyTorch memory, this process has 58.79 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 20.49 GiB is free. Including non-PyTorch memory, this process has 58.83 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 20.53 GiB is free. Including non-PyTorch memory, this process has 58.79 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 20.49 GiB is free. Including non-PyTorch memory, this process has 58.83 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 20.45 GiB is free. Including non-PyTorch memory, this process has 58.87 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 20.49 GiB is free. Including non-PyTorch memory, this process has 58.83 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 20.37 GiB is free. Including non-PyTorch memory, this process has 58.95 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 20.45 GiB is free. Including non-PyTorch memory, this process has 58.87 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 20.49 GiB is free. Including non-PyTorch memory, this process has 58.83 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 20.37 GiB is free. Including non-PyTorch memory, this process has 58.95 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 20.49 GiB is free. Including non-PyTorch memory, this process has 58.83 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 20.53 GiB is free. Including non-PyTorch memory, this process has 58.79 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items())result = OptimizerFromGradientAccumulator( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 20.41 GiB is free. Including non-PyTorch memory, this process has 58.91 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 20.49 GiB is free. Including non-PyTorch memory, this process has 58.83 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 20.53 GiB is free. Including non-PyTorch memory, this process has 58.79 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) +trainer = DistributedTrainer(config_file) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + + trainer = DistributedTrainer(config_file) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer(optimizer = ZeroDistributedOptimizer( + + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank)optimizer = optimizer_builder(param_groups_in_rank) + +optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator(result = OptimizerFromGradientAccumulator( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items())gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator(gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers(self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda")contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + +torch.cudatorch.cuda..OutOfMemoryErrorOutOfMemoryError: : CUDA out of memory. Tried to allocate 110.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 20.41 GiB is free. Including non-PyTorch memory, this process has 58.91 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFCUDA out of memory. Tried to allocate 110.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 20.53 GiB is free. Including non-PyTorch memory, this process has 58.79 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 20.49 GiB is free. Including non-PyTorch memory, this process has 58.83 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 110.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 20.53 GiB is free. Including non-PyTorch memory, this process has 58.79 GiB memory in use. Of the allocated memory 55.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 114.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 18.56 GiB is free. Including non-PyTorch memory, this process has 60.75 GiB memory in use. Of the allocated memory 57.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( +result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator(optimizer = optimizer_builder(param_groups_in_rank) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 114.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 18.56 GiB is free. Including non-PyTorch memory, this process has 60.75 GiB memory in use. Of the allocated memory 57.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 114.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 18.68 GiB is free. Including non-PyTorch memory, this process has 60.64 GiB memory in use. Of the allocated memory 57.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 114.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 18.45 GiB is free. Including non-PyTorch memory, this process has 60.87 GiB memory in use. Of the allocated memory 57.00 GiB is allocated by PyTorch, and 2.06 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[2024-12-27 17:40:57,538] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1132724 closing signal SIGTERM +[2024-12-27 17:40:57,539] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1132725 closing signal SIGTERM +[2024-12-27 17:40:57,539] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1132726 closing signal SIGTERM +[2024-12-27 17:40:57,539] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1132729 closing signal SIGTERM +[2024-12-27 17:40:57,539] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1132730 closing signal SIGTERM +[2024-12-27 17:40:57,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1632403 closing signal SIGTERM +[2024-12-27 17:40:57,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1632404 closing signal SIGTERM +[2024-12-27 17:40:57,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1632405 closing signal SIGTERM +[2024-12-27 17:40:57,544] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1632406 closing signal SIGTERM +[2024-12-27 17:40:57,544] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 3624149) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 17:40:57,546] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 679633 closing signal SIGTERM +[2024-12-27 17:40:57,546] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 679635 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:40:57 + host : ip-26-0-174-196.ec2.internal + rank : 17 (local_rank: 1) + exitcode : 1 (pid: 3624150) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_17:40:57 + host : ip-26-0-174-196.ec2.internal + rank : 18 (local_rank: 2) + exitcode : 1 (pid: 3624151) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_17:40:57 + host : ip-26-0-174-196.ec2.internal + rank : 19 (local_rank: 3) + exitcode : 1 (pid: 3624152) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-27_17:40:57 + host : ip-26-0-174-196.ec2.internal + rank : 20 (local_rank: 4) + exitcode : 1 (pid: 3624153) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-27_17:40:57 + host : ip-26-0-174-196.ec2.internal + rank : 21 (local_rank: 5) + exitcode : 1 (pid: 3624154) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-27_17:40:57 + host : ip-26-0-174-196.ec2.internal + rank : 22 (local_rank: 6) + exitcode : 1 (pid: 3624155) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-27_17:40:57 + host : ip-26-0-174-196.ec2.internal + rank : 23 (local_rank: 7) + exitcode : 1 (pid: 3624156) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:40:57 + host : ip-26-0-174-196.ec2.internal + rank : 16 (local_rank: 0) + exitcode : 1 (pid: 3624149) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-27 17:40:57,822] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1632401) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:40:57 + host : ip-26-0-168-95.ec2.internal + rank : 9 (local_rank: 1) + exitcode : 1 (pid: 1632402) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_17:40:57 + host : ip-26-0-168-95.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 1632407) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_17:40:57 + host : ip-26-0-168-95.ec2.internal + rank : 15 (local_rank: 7) + exitcode : 1 (pid: 1632408) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:40:57 + host : ip-26-0-168-95.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 1632401) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-27 17:40:57,916] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1132723) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:40:57 + host : ip-26-0-166-214.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 1132727) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_17:40:57 + host : ip-26-0-166-214.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 1132728) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:40:57 + host : ip-26-0-166-214.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 1132723) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-174-196: task 2: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13803433.0 +slurmstepd: error: *** STEP 13803433.0 ON ip-26-0-166-214 CANCELLED AT 2024-12-27T17:40:57 *** +[2024-12-27 17:40:57,976] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-27 17:40:57,976] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 679635 closing signal SIGTERM +srun: error: ip-26-0-168-95: task 1: Terminated +srun: error: ip-26-0-166-214: task 0: Terminated +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 679560 got signal: 15 +srun: error: ip-26-0-175-132: task 3: Exited with exit code 1 +srun: Force Terminated StepId=13803433.0 diff --git a/logs/13803461-bench_80G_dp2_tp4_pp4_acc4_mbs32_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13803461-bench_80G_dp2_tp4_pp4_acc4_mbs32_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..346fe09de428e0c400dcbcc8683033e259957443 --- /dev/null +++ b/logs/13803461-bench_80G_dp2_tp4_pp4_acc4_mbs32_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,9398 @@ ++ '[' -z 13803461 ']' ++ unset FI_PROVIDER ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-163-58,ip-26-0-164-75,ip-26-0-169-[139,207]' ++ export 'NODELIST=ip-26-0-163-58 +ip-26-0-164-75 +ip-26-0-169-139 +ip-26-0-169-207' ++ NODELIST='ip-26-0-163-58 +ip-26-0-164-75 +ip-26-0-169-139 +ip-26-0-169-207' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-163-58,ip-26-0-164-75,ip-26-0-169-[139,207]' ++ export MASTER_NODE=ip-26-0-163-58 ++ MASTER_NODE=ip-26-0-163-58 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ export NCCL_P2P_LEVEL=LOC ++ NCCL_P2P_LEVEL=LOC ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-163-58' +Master node: ip-26-0-163-58 ++ echo 'All nodes: ip-26-0-163-58 +ip-26-0-164-75 +ip-26-0-169-139 +ip-26-0-169-207' +All nodes: ip-26-0-163-58 +ip-26-0-164-75 +ip-26-0-169-139 +ip-26-0-169-207 ++ echo 'World size: 32' +World size: 32 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13803461 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-163-58:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_80G_dp2_tp4_pp4_acc4_mbs32_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-27 17:44:25,460] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:44:25,459] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:44:25,460] torch.distributed.run: [WARNING] +[2024-12-27 17:44:25,460] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:44:25,460] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:44:25,460] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:44:25,459] torch.distributed.run: [WARNING] +[2024-12-27 17:44:25,459] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:44:25,459] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:44:25,459] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:44:25,506] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:44:25,506] torch.distributed.run: [WARNING] +[2024-12-27 17:44:25,506] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:44:25,506] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:44:25,506] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:44:25,766] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:44:25,766] torch.distributed.run: [WARNING] +[2024-12-27 17:44:25,766] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:44:25,766] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:44:25,766] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-163-58:850785:850785 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-163-58:850785:850785 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:850785:850785 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-58:850785:850785 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:850788:850788 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850789:850789 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850787:850787 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850790:850790 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850786:850786 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850792:850792 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850791:850791 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:2115371:2115371 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:2115367:2115367 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:2115369:2115369 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:2115366:2115366 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:2115370:2115370 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:2115368:2115368 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:2115372:2115372 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:2115365:2115365 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850788:850788 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-163-58:850789:850789 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-163-58:850790:850790 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-163-58:850786:850786 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-164-75:1594323:1594323 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850787:850787 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-163-58:850788:850788 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:850789:850789 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:850788:850788 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-58:850789:850789 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-164-75:1594326:1594326 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-75:1594327:1594327 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-75:1594324:1594324 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-75:1594322:1594322 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-75:1594325:1594325 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-75:1594321:1594321 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-164-75:1594328:1594328 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850792:850792 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-163-58:850790:850790 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:850790:850790 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-58:850791:850791 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-163-58:850786:850786 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:850786:850786 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137566:2137566 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850787:850787 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:850787:850787 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137567:2137567 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850792:850792 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:850792:850792 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137563:2137563 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:850791:850791 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:850791:850791 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137562:2137562 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-139:2137564:2137564 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-139:2137565:2137565 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-139:2137569:2137569 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-139:2137568:2137568 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-207:2115371:2115371 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:2115369:2115369 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:2115366:2115366 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:2115367:2115367 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:2115370:2115370 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:2115365:2115365 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:2115368:2115368 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:2115372:2115372 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.169.207<0> +ip-26-0-169-207:2115371:2115371 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:2115371:2115371 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:2115369:2115369 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:2115367:2115367 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:2115370:2115370 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:2115366:2115366 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:2115369:2115369 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:2115370:2115370 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:2115367:2115367 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:2115366:2115366 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:2115365:2115365 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:2115365:2115365 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:2115372:2115372 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:2115372:2115372 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-207:2115368:2115368 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-207:2115368:2115368 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-164-75:1594326:1594326 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.164.75<0> +ip-26-0-164-75:1594327:1594327 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.164.75<0> +ip-26-0-164-75:1594323:1594323 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.164.75<0> +ip-26-0-164-75:1594328:1594328 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.164.75<0> +ip-26-0-164-75:1594325:1594325 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.164.75<0> +ip-26-0-164-75:1594324:1594324 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.164.75<0> +ip-26-0-164-75:1594321:1594321 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.164.75<0> +ip-26-0-164-75:1594322:1594322 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.164.75<0> +ip-26-0-169-139:2137567:2137567 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.169.139<0> +ip-26-0-169-139:2137566:2137566 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.169.139<0> +ip-26-0-169-139:2137563:2137563 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.169.139<0> +ip-26-0-169-139:2137562:2137562 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.169.139<0> +ip-26-0-164-75:1594325:1594325 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-75:1594324:1594324 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-75:1594322:1594322 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-75:1594327:1594327 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-75:1594325:1594325 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-164-75:1594324:1594324 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-164-75:1594322:1594322 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-164-75:1594327:1594327 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137564:2137564 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.169.139<0> +ip-26-0-164-75:1594326:1594326 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-75:1594326:1594326 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137565:2137565 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.169.139<0> +ip-26-0-169-139:2137569:2137569 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.169.139<0> +ip-26-0-169-139:2137568:2137568 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.169.139<0> +ip-26-0-164-75:1594328:1594328 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-75:1594328:1594328 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-164-75:1594323:1594323 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-75:1594323:1594323 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137567:2137567 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-139:2137567:2137567 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137562:2137562 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-139:2137562:2137562 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137566:2137566 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-139:2137563:2137563 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-139:2137566:2137566 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137563:2137563 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137565:2137565 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-139:2137565:2137565 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137568:2137568 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-139:2137568:2137568 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137564:2137564 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-139:2137564:2137564 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-164-75:1594321:1594321 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-164-75:1594321:1594321 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-139:2137569:2137569 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-139:2137569:2137569 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/253 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:850785:850947 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:850785:850947 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Using network Libfabric +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:850789:850949 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:850789:850949 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:850792:850948 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:850792:850948 [7] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Using network Libfabric +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:850786:850950 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:850786:850950 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:850788:850954 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:850788:850954 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:850790:850953 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:850790:850953 [5] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Using network Libfabric +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:850787:850951 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:850787:850951 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:850791:850952 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:850791:850952 [6] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO comm 0xa064e20 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-163-58:850785:850947 [0] NCCL INFO comm 0xa296e80 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-163-58:850787:850951 [2] NCCL INFO comm 0xa2e1960 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-163-58:850788:850954 [3] NCCL INFO comm 0xa142760 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-163-58:850789:850949 [4] NCCL INFO comm 0x9099270 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-163-58:850786:850950 [1] NCCL INFO comm 0x9728160 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-163-58:850791:850952 [6] NCCL INFO comm 0x8835210 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-163-58:850790:850953 [5] NCCL INFO comm 0x8f10230 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-163-58:850792:850948 [7] NCCL INFO comm 0x96f3ce0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO comm 0xa1633c0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO comm 0x906b2d0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO comm 0x86da5c0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO comm 0x9b4eaf0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO comm 0x984dd30 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO comm 0x8567180 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO comm 0x922fac0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO comm 0x99e5dd0 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO comm 0x9157cb0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO comm 0x96eaf30 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO comm 0xa415950 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO comm 0x961cb20 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO comm 0x96e11d0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO comm 0x8983840 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO comm 0x8717b20 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO comm 0x987e3a0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO comm 0x96e4ba0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO comm 0x8641a90 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO comm 0xa0b1250 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO comm 0x9ef44b0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO comm 0x88f9170 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO comm 0x9c1ecc0 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO comm 0x84d83a0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd48aa677614c5c9 - Init START +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-163-58:850792:850948 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850791:850952 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850790:850953 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850785:850947 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-58:850785:850947 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850786:850950 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-58:850786:850950 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:850790:850953 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-58:850790:850953 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:850791:850952 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-58:850791:850952 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850788:850954 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-58:850788:850954 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:850792:850948 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-58:850792:850948 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850789:850949 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-58:850789:850949 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850787:850951 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-58:850787:850951 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-163-58:850792:850948 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-163-58:850792:850948 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850791:850952 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-163-58:850791:850952 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850789:850949 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-163-58:850789:850949 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850787:850951 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-58:850787:850951 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Trees [0] 8/24/-1->16->0 [1] 8/24/-1->16->0 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850786:850950 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-58:850786:850950 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850788:850954 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-163-58:850788:850954 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850790:850953 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-163-58:850790:850953 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 15/-1/-1->31->-1 [3] 15/-1/-1->31->-1 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-58:850785:850947 [0] NCCL INFO Trees [0] 16/-1/-1->0->-1 [1] 16/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850785:850947 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:850791:850952 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:850789:850949 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:850792:850948 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:850787:850951 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:850786:850950 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:850788:850954 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:850790:850953 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:850785:850947 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Connected all rings +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Connected all rings +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Connected all rings +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Connected all rings +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Connected all rings +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Connected all rings +ip-26-0-163-58:850787:850951 [2] NCCL INFO Connected all rings +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Connected all rings +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Connected all rings +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Connected all rings +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Connected all rings +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Connected all rings +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Connected all rings +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Connected all rings +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Connected all rings +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Connected all rings +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Connected all rings +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Connected all rings +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Connected all rings +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Connected all rings +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Connected all rings +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Connected all rings +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Connected all rings +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Connected all rings +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Connected all rings +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Connected all rings +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Connected all rings +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Connected all rings +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Connected all rings +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Connected all rings +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Connected all rings +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:850948 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Connected all rings +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:850954 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850790:850953 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:850949 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:850952 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:850951 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:850950 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850785:850947 [0] NCCL INFO Connected all trees +ip-26-0-163-58:850785:850947 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850785:850947 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850785:850947 [0] NCCL INFO comm 0xa296e80 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO Connected all trees +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115372:2115529 [7] NCCL INFO comm 0xa064e20 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-163-58:850791:850952 [6] NCCL INFO Connected all trees +ip-26-0-163-58:850791:850952 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850791:850952 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850787:850951 [2] NCCL INFO Connected all trees +ip-26-0-163-58:850787:850951 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850787:850951 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850790:850953 [5] NCCL INFO Connected all trees +ip-26-0-163-58:850790:850953 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850790:850953 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850786:850950 [1] NCCL INFO Connected all trees +ip-26-0-163-58:850786:850950 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850786:850950 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO Connected all trees +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850791:850952 [6] NCCL INFO comm 0x8835210 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-163-58:850787:850951 [2] NCCL INFO comm 0xa2e1960 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO Connected all trees +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850790:850953 [5] NCCL INFO comm 0x8f10230 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO Connected all trees +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850786:850950 [1] NCCL INFO comm 0x9728160 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO Connected all trees +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO Connected all trees +ip-26-0-169-207:2115367:2115526 [2] NCCL INFO comm 0x9c1ecc0 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO Connected all trees +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO Connected all trees +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594327:1594480 [6] NCCL INFO comm 0x922fac0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO Connected all trees +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO Connected all trees +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115370:2115532 [5] NCCL INFO comm 0x961cb20 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-207:2115371:2115528 [6] NCCL INFO comm 0x88f9170 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO Connected all trees +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594326:1594484 [5] NCCL INFO comm 0x99e5dd0 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-207:2115366:2115530 [1] NCCL INFO comm 0x9ef44b0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-164-75:1594323:1594487 [2] NCCL INFO comm 0x9b4eaf0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-139:2137564:2137723 [2] NCCL INFO comm 0xa415950 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-164-75:1594322:1594485 [1] NCCL INFO comm 0x906b2d0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-139:2137568:2137728 [6] NCCL INFO comm 0x96e4ba0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO Connected all trees +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850792:850948 [7] NCCL INFO Connected all trees +ip-26-0-163-58:850792:850948 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850792:850948 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO Connected all trees +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850788:850954 [3] NCCL INFO Connected all trees +ip-26-0-163-58:850788:850954 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850788:850954 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137563:2137726 [1] NCCL INFO comm 0x96eaf30 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-163-58:850792:850948 [7] NCCL INFO comm 0x96f3ce0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-163-58:850789:850949 [4] NCCL INFO Connected all trees +ip-26-0-163-58:850789:850949 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850789:850949 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137567:2137725 [5] NCCL INFO comm 0x96e11d0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO Connected all trees +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850788:850954 [3] NCCL INFO comm 0xa142760 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-163-58:850789:850949 [4] NCCL INFO comm 0x9099270 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO Connected all trees +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115368:2115525 [3] NCCL INFO comm 0xa0b1250 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO Connected all trees +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115365:2115531 [0] NCCL INFO comm 0x8641a90 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO Connected all trees +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO Connected all trees +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115369:2115527 [4] NCCL INFO comm 0x84d83a0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-164-75:1594324:1594482 [3] NCCL INFO comm 0x86da5c0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO Connected all trees +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO Connected all trees +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594328:1594479 [7] NCCL INFO comm 0x8567180 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO Connected all trees +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO Connected all trees +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO Connected all trees +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137569:2137724 [7] NCCL INFO comm 0x8983840 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-139:2137566:2137727 [4] NCCL INFO comm 0x987e3a0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-164-75:1594325:1594483 [4] NCCL INFO comm 0x984dd30 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-139:2137565:2137722 [3] NCCL INFO comm 0x8717b20 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-164-75:1594321:1594486 [0] NCCL INFO comm 0xa1633c0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO Connected all trees +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137562:2137729 [0] NCCL INFO comm 0x9157cb0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfd48aa677614c5c9 - Init COMPLETE +ip-26-0-163-58:850785:851179 [0] NCCL INFO Using network Libfabric +ip-26-0-163-58:850788:851182 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:850791:851184 [6] NCCL INFO Using network Libfabric +ip-26-0-163-58:850787:851181 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:850790:851185 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:850786:851186 [1] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Using network Libfabric +ip-26-0-163-58:850789:851183 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:850792:851180 [7] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO comm 0x87a1a70 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO comm 0xa64fb40 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO comm 0x9927c80 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO comm 0x946c420 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO comm 0x9c22060 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO comm 0x9a893c0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO comm 0x89146b0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO comm 0x9d88a80 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-163-58:850792:851180 [7] NCCL INFO comm 0x992e2c0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-163-58:850791:851184 [6] NCCL INFO comm 0x8a6ec20 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-163-58:850790:851185 [5] NCCL INFO comm 0x914c740 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO comm 0x9393740 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO comm 0x92a4b40 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO comm 0x8952280 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO comm 0x9abacc0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO comm 0xa39f220 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-163-58:850788:851182 [3] NCCL INFO comm 0xa37c1c0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-163-58:850789:851183 [4] NCCL INFO comm 0x92d37b0 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-163-58:850787:851181 [2] NCCL INFO comm 0xa51bf30 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO comm 0x8bbeb80 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO comm 0xa2a0780 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO comm 0x8b33180 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-163-58:850786:851186 [1] NCCL INFO comm 0x9961bc0 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-163-58:850785:851179 [0] NCCL INFO comm 0xa4da060 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO comm 0x991f380 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO comm 0x991b580 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO comm 0x9856940 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO comm 0xa2eb400 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO comm 0xa12e3c0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO comm 0x887b770 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO comm 0x9e5aa80 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO comm 0x8712580 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x55e3399c0365d5f7 - Init START +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850791:851184 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850792:851180 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850790:851185 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850789:851183 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850788:851182 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850786:851186 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850785:851179 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-163-58:850787:851181 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850789:851183 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850786:851186 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:850792:851180 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:850787:851181 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-58:850791:851184 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850788:851182 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:850785:851179 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-58:850790:851185 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 15/-1/-1->31->-1 [3] 15/-1/-1->31->-1 +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850787:851181 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-58:850787:851181 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-58:850788:851182 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-163-58:850785:851179 [0] NCCL INFO Trees [0] 16/-1/-1->0->-1 [1] 16/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850788:851182 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850785:851179 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850791:851184 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-163-58:850786:851186 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Trees [0] 8/24/-1->16->0 [1] 8/24/-1->16->0 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-163-58:850791:851184 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850786:851186 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850790:851185 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-163-58:850792:851180 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-163-58:850790:851185 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850792:851180 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850789:851183 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-163-58:850789:851183 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Connected all rings +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Connected all rings +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Connected all rings +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Connected all rings +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Connected all rings +ip-26-0-163-58:850791:851184 [6] NCCL INFO Connected all rings +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Connected all rings +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Connected all rings +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Connected all rings +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Connected all rings +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Connected all rings +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Connected all rings +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Connected all rings +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Connected all rings +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Connected all rings +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Connected all rings +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Connected all rings +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Connected all rings +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Connected all rings +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Connected all rings +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Connected all rings +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Connected all rings +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Connected all rings +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Connected all rings +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Connected all rings +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Connected all rings +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Connected all rings +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Connected all rings +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Connected all rings +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Connected all rings +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Connected all rings +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851179 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Connected all rings +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:851183 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851184 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851181 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:851182 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851185 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850792:851180 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851186 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO Connected all trees +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115372:2115759 [7] NCCL INFO comm 0xa2a0780 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-163-58:850787:851181 [2] NCCL INFO Connected all trees +ip-26-0-163-58:850787:851181 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850787:851181 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850786:851186 [1] NCCL INFO Connected all trees +ip-26-0-163-58:850786:851186 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850786:851186 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO Connected all trees +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO Connected all trees +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO Connected all trees +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850787:851181 [2] NCCL INFO comm 0xa51bf30 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-163-58:850786:851186 [1] NCCL INFO comm 0x9961bc0 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-207:2115367:2115757 [2] NCCL INFO comm 0x9e5aa80 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO Connected all trees +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594327:1594715 [6] NCCL INFO comm 0x946c420 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO Connected all trees +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850785:851179 [0] NCCL INFO Connected all trees +ip-26-0-163-58:850785:851179 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850785:851179 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594323:1594714 [2] NCCL INFO comm 0x9d88a80 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO Connected all trees +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594322:1594711 [1] NCCL INFO comm 0x92a4b40 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-207:2115371:2115758 [6] NCCL INFO comm 0x8b33180 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-163-58:850785:851179 [0] NCCL INFO comm 0xa4da060 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO Connected all trees +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115370:2115755 [5] NCCL INFO comm 0x9856940 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-163-58:850791:851184 [6] NCCL INFO Connected all trees +ip-26-0-163-58:850791:851184 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850791:851184 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO Connected all trees +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO Connected all trees +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137564:2137957 [2] NCCL INFO comm 0xa64fb40 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-163-58:850791:851184 [6] NCCL INFO comm 0x8a6ec20 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-139:2137568:2137958 [6] NCCL INFO comm 0x991f380 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-163-58:850788:851182 [3] NCCL INFO Connected all trees +ip-26-0-163-58:850788:851182 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850788:851182 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137567:2137956 [5] NCCL INFO comm 0x991b580 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-163-58:850790:851185 [5] NCCL INFO Connected all trees +ip-26-0-163-58:850790:851185 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850790:851185 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850789:851183 [4] NCCL INFO Connected all trees +ip-26-0-163-58:850789:851183 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850789:851183 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850792:851180 [7] NCCL INFO Connected all trees +ip-26-0-163-58:850792:851180 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:850792:851180 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850788:851182 [3] NCCL INFO comm 0xa37c1c0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-163-58:850790:851185 [5] NCCL INFO comm 0x914c740 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO Connected all trees +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO Connected all trees +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850789:851183 [4] NCCL INFO comm 0x92d37b0 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-163-58:850792:851180 [7] NCCL INFO comm 0x992e2c0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-164-75:1594324:1594717 [3] NCCL INFO comm 0x89146b0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-164-75:1594328:1594718 [7] NCCL INFO comm 0x87a1a70 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO Connected all trees +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO Connected all trees +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO Connected all trees +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO Connected all trees +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO Connected all trees +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO Connected all trees +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115366:2115762 [1] NCCL INFO comm 0xa12e3c0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO Connected all trees +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO Connected all trees +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO Connected all trees +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594326:1594716 [5] NCCL INFO comm 0x9c22060 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-164-75:1594321:1594712 [0] NCCL INFO comm 0xa39f220 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-164-75:1594325:1594713 [4] NCCL INFO comm 0x9a893c0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-139:2137565:2137954 [3] NCCL INFO comm 0x8952280 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-207:2115368:2115760 [3] NCCL INFO comm 0xa2eb400 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO Connected all trees +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137569:2137959 [7] NCCL INFO comm 0x8bbeb80 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-139:2137563:2137955 [1] NCCL INFO comm 0x9927c80 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-139:2137566:2137953 [4] NCCL INFO comm 0x9abacc0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO Connected all trees +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115365:2115761 [0] NCCL INFO comm 0x887b770 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO Connected all trees +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115369:2115756 [4] NCCL INFO comm 0x8712580 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-169-139:2137562:2137960 [0] NCCL INFO comm 0x9393740 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x55e3399c0365d5f7 - Init COMPLETE +ip-26-0-163-58:850785:851213 [0] NCCL INFO Using network Libfabric +ip-26-0-163-58:850790:851220 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:850787:851218 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:850786:851219 [1] NCCL INFO Using network Libfabric +ip-26-0-163-58:850791:851215 [6] NCCL INFO Using network Libfabric +ip-26-0-163-58:850789:851216 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:850792:851214 [7] NCCL INFO Using network Libfabric +ip-26-0-163-58:850788:851217 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:850787:851218 [2] NCCL INFO comm 0xa52f560 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9972f3c7a1c3cba8 - Init START +ip-26-0-163-58:850786:851219 [1] NCCL INFO comm 0x99761d0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9972f3c7a1c3cba8 - Init START +ip-26-0-163-58:850785:851213 [0] NCCL INFO comm 0xa4ecd20 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9972f3c7a1c3cba8 - Init START +ip-26-0-163-58:850792:851214 [7] NCCL INFO comm 0x9941de0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9972f3c7a1c3cba8 - Init START +ip-26-0-163-58:850791:851215 [6] NCCL INFO comm 0x8a82e00 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9972f3c7a1c3cba8 - Init START +ip-26-0-163-58:850788:851217 [3] NCCL INFO comm 0xa390700 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9972f3c7a1c3cba8 - Init START +ip-26-0-163-58:850790:851220 [5] NCCL INFO comm 0x915f400 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9972f3c7a1c3cba8 - Init START +ip-26-0-163-58:850789:851216 [4] NCCL INFO comm 0x92e7790 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9972f3c7a1c3cba8 - Init START +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO comm 0x92b8e40 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3159b94a9664af80 - Init START +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO comm 0x9d9cd70 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3159b94a9664af80 - Init START +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO comm 0x8928690 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3159b94a9664af80 - Init START +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO comm 0xa3b1ee0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3159b94a9664af80 - Init START +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO comm 0x9a9c8c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x3159b94a9664af80 - Init START +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO comm 0x87b5b50 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3159b94a9664af80 - Init START +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO comm 0x947f0e0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x3159b94a9664af80 - Init START +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO comm 0x9c34d20 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3159b94a9664af80 - Init START +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO comm 0x93a6a50 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcbaaf56133722bf6 - Init START +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO comm 0x993a940 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xcbaaf56133722bf6 - Init START +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO comm 0x8bd2710 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcbaaf56133722bf6 - Init START +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO comm 0xa663e00 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xcbaaf56133722bf6 - Init START +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO comm 0x99332a0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcbaaf56133722bf6 - Init START +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO comm 0x8966a00 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xcbaaf56133722bf6 - Init START +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO comm 0x9acd980 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcbaaf56133722bf6 - Init START +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO comm 0x992f780 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcbaaf56133722bf6 - Init START +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO comm 0xa1425c0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x476eaeeab2e93e98 - Init START +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO comm 0x8890a90 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x476eaeeab2e93e98 - Init START +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO comm 0x8b46fc0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x476eaeeab2e93e98 - Init START +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO comm 0xa2b3440 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x476eaeeab2e93e98 - Init START +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO comm 0x986aac0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x476eaeeab2e93e98 - Init START +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO comm 0x9e6d740 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x476eaeeab2e93e98 - Init START +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO comm 0x87268c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x476eaeeab2e93e98 - Init START +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO comm 0xa2ffa50 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x476eaeeab2e93e98 - Init START +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850786:851219 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850787:851218 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850789:851216 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850785:851213 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-163-58:850792:851214 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850788:851217 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850791:851215 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850790:851220 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:850788:851217 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-58:850792:851214 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850785:851213 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850787:851218 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850791:851215 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850790:851220 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Connected all rings +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Connected all rings +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Connected all rings +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Connected all rings +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-163-58:850789:851216 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850791:851215 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-163-58:850791:851215 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850790:851220 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-163-58:850790:851220 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850788:851217 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-163-58:850792:851214 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-163-58:850787:851218 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-58:850788:851217 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850792:851214 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850787:851218 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850786:851219 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-58:850786:851219 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:850785:851213 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850785:851213 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Connected all rings +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Connected all rings +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Connected all rings +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Connected all rings +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Connected all rings +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Connected all rings +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Connected all rings +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Connected all rings +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Connected all rings +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Connected all rings +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Connected all rings +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Connected all rings +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO Connected all trees +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115372:2115792 [7] NCCL INFO comm 0xa2b3440 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x476eaeeab2e93e98 - Init COMPLETE +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO Connected all trees +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO Connected all trees +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137562:2137987 [0] NCCL INFO comm 0x93a6a50 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcbaaf56133722bf6 - Init COMPLETE +ip-26-0-169-207:2115365:2115788 [0] NCCL INFO comm 0x8890a90 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x476eaeeab2e93e98 - Init COMPLETE +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO Connected all trees +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO Connected all trees +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115366:2115793 [1] NCCL INFO comm 0xa1425c0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x476eaeeab2e93e98 - Init COMPLETE +ip-26-0-169-139:2137569:2137990 [7] NCCL INFO comm 0x8bd2710 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcbaaf56133722bf6 - Init COMPLETE +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO Connected all trees +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115367:2115794 [2] NCCL INFO comm 0x9e6d740 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x476eaeeab2e93e98 - Init COMPLETE +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO Connected all trees +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO Connected all trees +ip-26-0-169-207:2115371:2115789 [6] NCCL INFO comm 0x8b46fc0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x476eaeeab2e93e98 - Init COMPLETE +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO Connected all trees +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO Connected all trees +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO Connected all trees +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115368:2115795 [3] NCCL INFO comm 0xa2ffa50 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x476eaeeab2e93e98 - Init COMPLETE +ip-26-0-169-139:2137564:2137991 [2] NCCL INFO comm 0xa663e00 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xcbaaf56133722bf6 - Init COMPLETE +ip-26-0-169-139:2137563:2137989 [1] NCCL INFO comm 0x993a940 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xcbaaf56133722bf6 - Init COMPLETE +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO Connected all trees +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137568:2137993 [6] NCCL INFO comm 0x99332a0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcbaaf56133722bf6 - Init COMPLETE +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO Connected all trees +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO Connected all trees +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137567:2137988 [5] NCCL INFO comm 0x992f780 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcbaaf56133722bf6 - Init COMPLETE +ip-26-0-169-207:2115370:2115790 [5] NCCL INFO comm 0x986aac0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x476eaeeab2e93e98 - Init COMPLETE +ip-26-0-169-207:2115369:2115791 [4] NCCL INFO comm 0x87268c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x476eaeeab2e93e98 - Init COMPLETE +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Connected all rings +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Connected all rings +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Connected all rings +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Connected all rings +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Connected all rings +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Connected all rings +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Connected all rings +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Connected all rings +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO Connected all trees +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850788:851217 [3] NCCL INFO Connected all rings +ip-26-0-163-58:850790:851220 [5] NCCL INFO Connected all rings +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Connected all rings +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO Connected all trees +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850789:851216 [4] NCCL INFO Connected all rings +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Connected all rings +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850792:851214 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850788:851217 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851220 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850786:851219 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Connected all rings +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2137992 [3] NCCL INFO comm 0x8966a00 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xcbaaf56133722bf6 - Init COMPLETE +ip-26-0-163-58:850787:851218 [2] NCCL INFO Connected all rings +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Connected all rings +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2137994 [4] NCCL INFO comm 0x9acd980 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcbaaf56133722bf6 - Init COMPLETE +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:851213 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:851216 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851215 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851218 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO Connected all trees +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594328:1594748 [7] NCCL INFO comm 0x87b5b50 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3159b94a9664af80 - Init COMPLETE +ip-26-0-163-58:850785:851213 [0] NCCL INFO Connected all trees +ip-26-0-163-58:850785:851213 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:850785:851213 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850785:851213 [0] NCCL INFO comm 0xa4ecd20 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9972f3c7a1c3cba8 - Init COMPLETE +ip-26-0-163-58:850792:851214 [7] NCCL INFO Connected all trees +ip-26-0-163-58:850792:851214 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:850792:851214 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO Connected all trees +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850792:851214 [7] NCCL INFO comm 0x9941de0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9972f3c7a1c3cba8 - Init COMPLETE +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO Connected all trees +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594321:1594744 [0] NCCL INFO comm 0xa3b1ee0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3159b94a9664af80 - Init COMPLETE +ip-26-0-164-75:1594322:1594747 [1] NCCL INFO comm 0x92b8e40 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3159b94a9664af80 - Init COMPLETE +ip-26-0-163-58:850786:851219 [1] NCCL INFO Connected all trees +ip-26-0-163-58:850786:851219 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:850786:851219 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO Connected all trees +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO Connected all trees +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850786:851219 [1] NCCL INFO comm 0x99761d0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9972f3c7a1c3cba8 - Init COMPLETE +ip-26-0-163-58:850791:851215 [6] NCCL INFO Connected all trees +ip-26-0-163-58:850791:851215 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:850791:851215 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594327:1594746 [6] NCCL INFO comm 0x947f0e0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x3159b94a9664af80 - Init COMPLETE +ip-26-0-164-75:1594323:1594750 [2] NCCL INFO comm 0x9d9cd70 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3159b94a9664af80 - Init COMPLETE +ip-26-0-163-58:850791:851215 [6] NCCL INFO comm 0x8a82e00 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9972f3c7a1c3cba8 - Init COMPLETE +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO Connected all trees +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594324:1594745 [3] NCCL INFO comm 0x8928690 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3159b94a9664af80 - Init COMPLETE +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO Connected all trees +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO Connected all trees +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850790:851220 [5] NCCL INFO Connected all trees +ip-26-0-163-58:850790:851220 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:850790:851220 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850787:851218 [2] NCCL INFO Connected all trees +ip-26-0-163-58:850787:851218 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:850787:851218 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850788:851217 [3] NCCL INFO Connected all trees +ip-26-0-163-58:850788:851217 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:850788:851217 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850789:851216 [4] NCCL INFO Connected all trees +ip-26-0-163-58:850789:851216 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:850789:851216 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594326:1594751 [5] NCCL INFO comm 0x9c34d20 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3159b94a9664af80 - Init COMPLETE +ip-26-0-164-75:1594325:1594749 [4] NCCL INFO comm 0x9a9c8c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x3159b94a9664af80 - Init COMPLETE +ip-26-0-163-58:850790:851220 [5] NCCL INFO comm 0x915f400 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9972f3c7a1c3cba8 - Init COMPLETE +ip-26-0-163-58:850787:851218 [2] NCCL INFO comm 0xa52f560 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9972f3c7a1c3cba8 - Init COMPLETE +ip-26-0-163-58:850788:851217 [3] NCCL INFO comm 0xa390700 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9972f3c7a1c3cba8 - Init COMPLETE +ip-26-0-163-58:850789:851216 [4] NCCL INFO comm 0x92e7790 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9972f3c7a1c3cba8 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:850791:851255 [6] NCCL INFO Using network Libfabric +ip-26-0-163-58:850789:851254 [4] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:850790:851257 [5] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Using network Libfabric +ip-26-0-163-58:850792:851256 [7] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO comm 0x9eb7470 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xecf0cf753057f681 - Init START +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO comm 0xac1aa80 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xecf0cf753057f681 - Init START +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO comm 0xac14fb0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xecf0cf753057f681 - Init START +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO comm 0xb7facb0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xecf0cf753057f681 - Init START +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO comm 0xac20610 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x10253e1b8202ef25 - Init START +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO comm 0xb94c6d0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x10253e1b8202ef25 - Init START +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO comm 0xa68cf60 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x10253e1b8202ef25 - Init START +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO comm 0x9c4cd00 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x10253e1b8202ef25 - Init START +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO comm 0xbb975b0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf3e106eb3ebdf004 - Init START +ip-26-0-163-58:850792:851256 [7] NCCL INFO comm 0xadee1f0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9d505f7997ca0af1 - Init START +ip-26-0-163-58:850789:851254 [4] NCCL INFO comm 0xa7935a0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9d505f7997ca0af1 - Init START +ip-26-0-163-58:850791:851255 [6] NCCL INFO comm 0x9f2d330 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9d505f7997ca0af1 - Init START +ip-26-0-163-58:850790:851257 [5] NCCL INFO comm 0xa60b530 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9d505f7997ca0af1 - Init START +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO comm 0x9b76320 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf3e106eb3ebdf004 - Init START +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO comm 0xc029c80 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf3e106eb3ebdf004 - Init START +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO comm 0xb4281c0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf3e106eb3ebdf004 - Init START +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO comm 0x9a0ce40 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x12a291f08637c3b - Init START +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO comm 0xab50c20 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x12a291f08637c3b - Init START +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO comm 0xbfdc620 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x12a291f08637c3b - Init START +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO comm 0x9e2c680 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x12a291f08637c3b - Init START +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO comm 0xa59e740 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7e103957b592facf - Init START +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO comm 0xb6984c0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7e103957b592facf - Init START +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO comm 0xb083050 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7e103957b592facf - Init START +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO comm 0x9c0d370 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7e103957b592facf - Init START +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850789:851254 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850792:851256 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850791:851255 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850790:851257 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO comm 0xb95dca0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x41765d6e8810f0c2 - Init START +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO comm 0xa764450 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x41765d6e8810f0c2 - Init START +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO comm 0xad82cd0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x41765d6e8810f0c2 - Init START +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO comm 0xa4e1ce0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x41765d6e8810f0c2 - Init START +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850785:851259 [0] NCCL INFO Using network Libfabric +ip-26-0-163-58:850786:851261 [1] NCCL INFO Using network Libfabric +ip-26-0-163-58:850788:851262 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:850787:851260 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:850788:851262 [3] NCCL INFO comm 0xb83b480 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb490703616ea2195 - Init START +ip-26-0-163-58:850787:851260 [2] NCCL INFO comm 0xb9d9890 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb490703616ea2195 - Init START +ip-26-0-163-58:850785:851259 [0] NCCL INFO comm 0xb994a00 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb490703616ea2195 - Init START +ip-26-0-163-58:850786:851261 [1] NCCL INFO comm 0xae20860 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb490703616ea2195 - Init START +ip-26-0-163-58:850786:851261 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850785:851259 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-163-58:850788:851262 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850787:851260 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:850788:851262 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850791:851255 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:850785:851259 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850789:851254 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850790:851257 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850786:851261 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-163-58:850791:851255 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-163-58:850792:851256 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-58:850791:851255 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-163-58:850792:851256 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-58:850790:851257 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-58:850789:851254 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850790:851257 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850789:851254 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850787:851260 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-58:850787:851260 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-58:850787:851260 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850788:851262 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-58:850788:851262 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850786:851261 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-58:850786:851261 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-58:850785:851259 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850785:851259 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Connected all rings +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Connected all rings +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Connected all rings +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Connected all rings +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Connected all rings +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Connected all rings +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Connected all rings +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Connected all rings +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Connected all rings +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Connected all rings +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Connected all rings +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Connected all rings +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Connected all rings +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850789:851254 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851255 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Connected all rings +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851257 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Connected all rings +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Connected all rings +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851261 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Connected all rings +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850785:851259 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851260 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Connected all rings +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Connected all rings +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Connected all rings +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Connected all rings +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Connected all rings +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Connected all rings +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Connected all rings +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO Connected all trees +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115365:2115831 [0] NCCL INFO comm 0x9b76320 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf3e106eb3ebdf004 - Init COMPLETE +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO Connected all trees +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO Connected all trees +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115372:2115838 [7] NCCL INFO comm 0xbfdc620 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x12a291f08637c3b - Init COMPLETE +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Connected all rings +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Connected all rings +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115369:2115833 [4] NCCL INFO comm 0x9a0ce40 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x12a291f08637c3b - Init COMPLETE +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Connected all rings +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO Connected all trees +ip-26-0-163-58:850792:851256 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850792:851256 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Connected all rings +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Connected all rings +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Connected all rings +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850788:851262 [3] NCCL INFO Connected all trees +ip-26-0-163-58:850788:851262 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850788:851262 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Connected all rings +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Connected all rings +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO Connected all trees +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850792:851256 [7] NCCL INFO comm 0xadee1f0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9d505f7997ca0af1 - Init COMPLETE +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO Connected all trees +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO Connected all trees +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850788:851262 [3] NCCL INFO comm 0xb83b480 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb490703616ea2195 - Init COMPLETE +ip-26-0-169-207:2115368:2115835 [3] NCCL INFO comm 0xc029c80 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf3e106eb3ebdf004 - Init COMPLETE +ip-26-0-163-58:850785:851259 [0] NCCL INFO Connected all trees +ip-26-0-163-58:850785:851259 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850785:851259 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115367:2115834 [2] NCCL INFO comm 0xbb975b0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf3e106eb3ebdf004 - Init COMPLETE +ip-26-0-169-207:2115366:2115837 [1] NCCL INFO comm 0xb4281c0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf3e106eb3ebdf004 - Init COMPLETE +ip-26-0-163-58:850785:851259 [0] NCCL INFO comm 0xb994a00 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb490703616ea2195 - Init COMPLETE +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO Connected all trees +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO Connected all trees +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO Connected all trees +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850789:851254 [4] NCCL INFO Connected all trees +ip-26-0-163-58:850789:851254 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850789:851254 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594324:1594793 [3] NCCL INFO comm 0x9c0d370 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7e103957b592facf - Init COMPLETE +ip-26-0-169-207:2115371:2115839 [6] NCCL INFO comm 0x9e2c680 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x12a291f08637c3b - Init COMPLETE +ip-26-0-169-207:2115370:2115836 [5] NCCL INFO comm 0xab50c20 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x12a291f08637c3b - Init COMPLETE +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO Connected all trees +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850789:851254 [4] NCCL INFO comm 0xa7935a0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9d505f7997ca0af1 - Init COMPLETE +ip-26-0-164-75:1594325:1594787 [4] NCCL INFO comm 0xad82cd0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x41765d6e8810f0c2 - Init COMPLETE +ip-26-0-163-58:850786:851261 [1] NCCL INFO Connected all trees +ip-26-0-163-58:850786:851261 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850786:851261 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850787:851260 [2] NCCL INFO Connected all trees +ip-26-0-163-58:850787:851260 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850787:851260 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850791:851255 [6] NCCL INFO Connected all trees +ip-26-0-163-58:850791:851255 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850791:851255 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850790:851257 [5] NCCL INFO Connected all trees +ip-26-0-163-58:850790:851257 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850790:851257 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850786:851261 [1] NCCL INFO comm 0xae20860 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb490703616ea2195 - Init COMPLETE +ip-26-0-163-58:850787:851260 [2] NCCL INFO comm 0xb9d9890 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb490703616ea2195 - Init COMPLETE +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO Connected all trees +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850791:851255 [6] NCCL INFO comm 0x9f2d330 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x9d505f7997ca0af1 - Init COMPLETE +ip-26-0-163-58:850790:851257 [5] NCCL INFO comm 0xa60b530 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9d505f7997ca0af1 - Init COMPLETE +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO Connected all trees +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594321:1594786 [0] NCCL INFO comm 0xb6984c0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7e103957b592facf - Init COMPLETE +ip-26-0-164-75:1594328:1594792 [7] NCCL INFO comm 0xa4e1ce0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x41765d6e8810f0c2 - Init COMPLETE +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO Connected all trees +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO Connected all trees +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO Connected all trees +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO Connected all trees +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594322:1594789 [1] NCCL INFO comm 0xa59e740 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7e103957b592facf - Init COMPLETE +ip-26-0-164-75:1594327:1594791 [6] NCCL INFO comm 0xa764450 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x41765d6e8810f0c2 - Init COMPLETE +ip-26-0-164-75:1594323:1594788 [2] NCCL INFO comm 0xb083050 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7e103957b592facf - Init COMPLETE +ip-26-0-164-75:1594326:1594790 [5] NCCL INFO comm 0xb95dca0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x41765d6e8810f0c2 - Init COMPLETE +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO Connected all trees +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO Connected all trees +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137565:2138033 [3] NCCL INFO comm 0x9c4cd00 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x10253e1b8202ef25 - Init COMPLETE +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO Connected all trees +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:850788:851284 [3] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO comm 0xc044f90 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x82f7c2878f53274f - Init START +ip-26-0-163-58:850788:851284 [3] NCCL INFO comm 0xc27f9a0 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x82f7c2878f53274f - Init START +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO comm 0xa652130 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x82f7c2878f53274f - Init START +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO comm 0xa691a90 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x82f7c2878f53274f - Init START +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850788:851284 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO Connected all trees +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137562:2138029 [0] NCCL INFO comm 0xa68cf60 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x10253e1b8202ef25 - Init COMPLETE +ip-26-0-169-139:2137569:2138030 [7] NCCL INFO comm 0x9eb7470 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xecf0cf753057f681 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:850792:851287 [7] NCCL INFO Using network Libfabric +ip-26-0-163-58:850785:851288 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO comm 0xbff6f50 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcc46a5f6a495e5ee - Init START +ip-26-0-163-58:850792:851287 [7] NCCL INFO comm 0xb832a20 rank 0 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcc46a5f6a495e5ee - Init START +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO comm 0xa4fcf10 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcc46a5f6a495e5ee - Init START +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO comm 0xa8fc250 rank 2 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcc46a5f6a495e5ee - Init START +ip-26-0-163-58:850792:851287 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850785:851288 [0] NCCL INFO comm 0xc3d9390 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2aa9da01a75974e2 - Init START +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO comm 0xc0dc960 rank 1 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2aa9da01a75974e2 - Init START +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO comm 0xa5baa00 rank 3 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2aa9da01a75974e2 - Init START +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO comm 0xb0d1bc0 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2aa9da01a75974e2 - Init START +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850785:851288 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137566:2138028 [4] NCCL INFO comm 0xb7facb0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xecf0cf753057f681 - Init COMPLETE +ip-26-0-163-58:850789:851290 [4] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO comm 0xa451750 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8710b76650e401f1 - Init START +ip-26-0-163-58:850789:851290 [4] NCCL INFO comm 0xb1d7a00 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8710b76650e401f1 - Init START +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO comm 0xb815f90 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8710b76650e401f1 - Init START +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO comm 0xb7c79f0 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8710b76650e401f1 - Init START +ip-26-0-163-58:850789:851290 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850788:851284 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:850792:851287 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:850789:851290 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-58:850788:851284 [3] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850788:851284 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850785:851288 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-58:850785:851288 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850785:851288 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-58:850792:851287 [7] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850792:851287 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-58:850789:851290 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850789:851290 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO Connected all trees +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO Connected all trees +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137563:2138034 [1] NCCL INFO comm 0xac20610 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x10253e1b8202ef25 - Init COMPLETE +ip-26-0-169-139:2137564:2138035 [2] NCCL INFO comm 0xb94c6d0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x10253e1b8202ef25 - Init COMPLETE +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO Connected all trees +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:850786:851300 [1] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:850787:851302 [2] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO comm 0xbe6ca40 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2ce709210da40bab - Init START +ip-26-0-163-58:850786:851300 [1] NCCL INFO comm 0xb865170 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2ce709210da40bab - Init START +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO comm 0xafe36c0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2ce709210da40bab - Init START +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO comm 0xb665080 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2ce709210da40bab - Init START +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850786:851300 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO Connected all trees +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO comm 0xbbb1ab0 rank 3 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a7a083ee6ff7287 - Init START +ip-26-0-163-58:850787:851302 [2] NCCL INFO comm 0xc41e280 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a7a083ee6ff7287 - Init START +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO comm 0xbac79b0 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a7a083ee6ff7287 - Init START +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO comm 0xc391040 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a7a083ee6ff7287 - Init START +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850787:851302 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137568:2138032 [6] NCCL INFO comm 0xac1aa80 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xecf0cf753057f681 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:850791:851304 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO comm 0xa870f60 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbc2cfe7c1505c6e5 - Init START +ip-26-0-163-58:850791:851304 [6] NCCL INFO comm 0xa9715f0 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbc2cfe7c1505c6e5 - Init START +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO comm 0xb1a9060 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbc2cfe7c1505c6e5 - Init START +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO comm 0xb65ebf0 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbc2cfe7c1505c6e5 - Init START +ip-26-0-169-139:2137567:2138031 [5] NCCL INFO comm 0xac14fb0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xecf0cf753057f681 - Init COMPLETE +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850791:851304 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:850790:851307 [5] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO comm 0xb5950a0 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x84e06f0fbba57a1d - Init START +ip-26-0-163-58:850790:851307 [5] NCCL INFO comm 0xb04fb30 rank 0 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x84e06f0fbba57a1d - Init START +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO comm 0xb977500 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x84e06f0fbba57a1d - Init START +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO comm 0xb6597b0 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x84e06f0fbba57a1d - Init START +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850790:851307 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 00/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 01/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 02/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 03/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 00/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 00/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 01/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 01/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 02/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 02/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 03/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 03/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 00/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 01/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 02/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 03/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 00/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 01/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 02/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 00/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 03/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 01/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 02/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 03/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 00/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 01/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 02/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 03/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 00/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 01/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 02/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 03/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:850787:851302 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:850786:851300 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-58:850787:851302 [2] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850787:851302 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850786:851300 [1] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850786:851300 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850790:851307 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850791:851304 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-58:850790:851307 [5] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850790:851307 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-58:850791:851304 [6] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850791:851304 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 00/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 01/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 02/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 00/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 03/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 01/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 02/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 03/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 00/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 01/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 02/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 03/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 00/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 01/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 02/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 03/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Connected all rings +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Connected all rings +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Connected all rings +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Connected all rings +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Connected all rings +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 02/0 : 1[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 03/0 : 1[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 02/0 : 3[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 03/0 : 3[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Connected all rings +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851290 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Connected all rings +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851284 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Connected all rings +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Connected all rings +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Connected all rings +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Connected all rings +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 00/0 : 0[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 00/0 : 2[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 01/0 : 0[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 00/0 : 2[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 01/0 : 2[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 01/0 : 2[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 00/0 : 0[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 01/0 : 0[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851287 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 00/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 01/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 02/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Channel 03/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Connected all rings +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 00/0 : 3[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851288 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 01/0 : 3[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Connected all rings +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Connected all rings +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Connected all rings +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 00/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 01/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 02/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 03/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Connected all rings +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 02/0 : 3[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 03/0 : 3[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 02/0 : 1[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 03/0 : 1[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 00/0 : 3[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Channel 01/0 : 3[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO Connected all trees +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850788:851284 [3] NCCL INFO Connected all trees +ip-26-0-163-58:850788:851284 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850788:851284 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO Connected all trees +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850792:851287 [7] NCCL INFO Connected all trees +ip-26-0-163-58:850792:851287 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850792:851287 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO Connected all trees +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115369:2115863 [4] NCCL INFO comm 0xa451750 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8710b76650e401f1 - Init COMPLETE +ip-26-0-163-58:850788:851284 [3] NCCL INFO comm 0xc27f9a0 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x82f7c2878f53274f - Init COMPLETE +ip-26-0-169-207:2115372:2115861 [7] NCCL INFO comm 0xbff6f50 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcc46a5f6a495e5ee - Init COMPLETE +ip-26-0-163-58:850792:851287 [7] NCCL INFO comm 0xb832a20 rank 0 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcc46a5f6a495e5ee - Init COMPLETE +ip-26-0-169-207:2115368:2115860 [3] NCCL INFO comm 0xc044f90 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x82f7c2878f53274f - Init COMPLETE +ip-26-0-163-58:850789:851290 [4] NCCL INFO Connected all trees +ip-26-0-163-58:850789:851290 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850789:851290 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO Connected all trees +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850785:851288 [0] NCCL INFO Connected all trees +ip-26-0-163-58:850785:851288 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850785:851288 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850789:851290 [4] NCCL INFO comm 0xb1d7a00 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8710b76650e401f1 - Init COMPLETE +ip-26-0-169-207:2115365:2115862 [0] NCCL INFO comm 0xa5baa00 rank 3 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2aa9da01a75974e2 - Init COMPLETE +ip-26-0-163-58:850785:851288 [0] NCCL INFO comm 0xc3d9390 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2aa9da01a75974e2 - Init COMPLETE +ip-26-0-163-58:850787:851302 [2] NCCL INFO Connected all rings +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Connected all rings +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Connected all rings +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 00/0 : 2[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 01/0 : 2[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 00/0 : 0[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 01/0 : 0[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Connected all rings +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 02/0 : 3[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 03/0 : 3[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Connected all rings +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Connected all rings +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Connected all rings +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Connected all rings +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 02/0 : 3[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 03/0 : 3[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 00/0 : 3[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Channel 01/0 : 3[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Connected all rings +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Connected all rings +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 00/0 : 3[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Channel 01/0 : 3[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO Connected all trees +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO Connected all trees +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO Connected all trees +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Connected all rings +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO Connected all trees +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO Connected all trees +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO Connected all trees +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Connected all rings +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Connected all rings +ip-26-0-169-139:2137569:2138058 [7] NCCL INFO comm 0xa8fc250 rank 2 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcc46a5f6a495e5ee - Init COMPLETE +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 00/0 : 3[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 01/0 : 3[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851302 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594328:1594816 [7] NCCL INFO comm 0xa4fcf10 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcc46a5f6a495e5ee - Init COMPLETE +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Connected all rings +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594325:1594818 [4] NCCL INFO comm 0xb7c79f0 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8710b76650e401f1 - Init COMPLETE +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 00/0 : 3[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 01/0 : 3[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851300 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137565:2138057 [3] NCCL INFO comm 0xa691a90 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x82f7c2878f53274f - Init COMPLETE +ip-26-0-164-75:1594324:1594815 [3] NCCL INFO comm 0xa652130 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x82f7c2878f53274f - Init COMPLETE +ip-26-0-169-139:2137566:2138060 [4] NCCL INFO comm 0xb815f90 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8710b76650e401f1 - Init COMPLETE +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO Connected all trees +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO Connected all trees +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Connected all rings +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 00/0 : 0[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 01/0 : 0[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 00/0 : 2[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 01/0 : 2[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Connected all rings +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851304 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851307 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594321:1594817 [0] NCCL INFO comm 0xc0dc960 rank 1 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2aa9da01a75974e2 - Init COMPLETE +ip-26-0-169-139:2137562:2138059 [0] NCCL INFO comm 0xb0d1bc0 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2aa9da01a75974e2 - Init COMPLETE +ip-26-0-163-58:850787:851302 [2] NCCL INFO Connected all trees +ip-26-0-163-58:850787:851302 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850787:851302 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850786:851300 [1] NCCL INFO Connected all trees +ip-26-0-163-58:850786:851300 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850786:851300 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850787:851302 [2] NCCL INFO comm 0xc41e280 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a7a083ee6ff7287 - Init COMPLETE +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO Connected all trees +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO Connected all trees +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850786:851300 [1] NCCL INFO comm 0xb865170 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2ce709210da40bab - Init COMPLETE +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO Connected all trees +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850791:851304 [6] NCCL INFO Connected all trees +ip-26-0-163-58:850791:851304 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850791:851304 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115367:2115873 [2] NCCL INFO comm 0xbbb1ab0 rank 3 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a7a083ee6ff7287 - Init COMPLETE +ip-26-0-169-207:2115366:2115872 [1] NCCL INFO comm 0xbe6ca40 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2ce709210da40bab - Init COMPLETE +ip-26-0-169-207:2115370:2115875 [5] NCCL INFO comm 0xb5950a0 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x84e06f0fbba57a1d - Init COMPLETE +ip-26-0-163-58:850791:851304 [6] NCCL INFO comm 0xa9715f0 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbc2cfe7c1505c6e5 - Init COMPLETE +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO Connected all trees +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850790:851307 [5] NCCL INFO Connected all trees +ip-26-0-163-58:850790:851307 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:850790:851307 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115371:2115874 [6] NCCL INFO comm 0xa870f60 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbc2cfe7c1505c6e5 - Init COMPLETE +ip-26-0-163-58:850790:851307 [5] NCCL INFO comm 0xb04fb30 rank 0 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x84e06f0fbba57a1d - Init COMPLETE +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO Connected all trees +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO Connected all trees +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO Connected all trees +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO Connected all trees +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO Connected all trees +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137563:2138066 [1] NCCL INFO comm 0xb665080 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2ce709210da40bab - Init COMPLETE +ip-26-0-164-75:1594322:1594827 [1] NCCL INFO comm 0xafe36c0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2ce709210da40bab - Init COMPLETE +ip-26-0-169-139:2137564:2138067 [2] NCCL INFO comm 0xc391040 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a7a083ee6ff7287 - Init COMPLETE +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO Connected all trees +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594323:1594828 [2] NCCL INFO comm 0xbac79b0 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1a7a083ee6ff7287 - Init COMPLETE +ip-26-0-164-75:1594326:1594830 [5] NCCL INFO comm 0xb977500 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x84e06f0fbba57a1d - Init COMPLETE +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO Connected all trees +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO Connected all trees +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594327:1594829 [6] NCCL INFO comm 0xb1a9060 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbc2cfe7c1505c6e5 - Init COMPLETE +ip-26-0-169-139:2137567:2138070 [5] NCCL INFO comm 0xb6597b0 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x84e06f0fbba57a1d - Init COMPLETE +ip-26-0-169-139:2137568:2138069 [6] NCCL INFO comm 0xb65ebf0 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbc2cfe7c1505c6e5 - Init COMPLETE +ip-26-0-163-58:850785:851317 [0] NCCL INFO Using network Libfabric +ip-26-0-163-58:850789:851318 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:850789:851318 [4] NCCL INFO comm 0xb205350 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8863d0fafd732fbd - Init START +ip-26-0-163-58:850785:851317 [0] NCCL INFO comm 0xc407f70 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8863d0fafd732fbd - Init START +ip-26-0-163-58:850789:851318 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850785:851317 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-163-58:850786:851320 [1] NCCL INFO Using network Libfabric +ip-26-0-163-58:850790:851321 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:850786:851320 [1] NCCL INFO comm 0xb892430 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x32efecc46f3158e9 - Init START +ip-26-0-163-58:850790:851321 [5] NCCL INFO comm 0xb07be60 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x32efecc46f3158e9 - Init START +ip-26-0-163-58:850786:851320 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850790:851321 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850792:851324 [7] NCCL INFO Using network Libfabric +ip-26-0-163-58:850788:851323 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:850788:851323 [3] NCCL INFO comm 0xc2abdd0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x53d3582ba4770597 - Init START +ip-26-0-163-58:850792:851324 [7] NCCL INFO comm 0xb85f170 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x53d3582ba4770597 - Init START +ip-26-0-163-58:850792:851324 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850788:851323 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO comm 0xb836490 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4beee5f9ade2bcab - Init START +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO comm 0xb10ce20 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4beee5f9ade2bcab - Init START +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850787:851326 [2] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:850791:851327 [6] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO comm 0xc1188e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe8409e998f1197a1 - Init START +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO comm 0xb802f00 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe8409e998f1197a1 - Init START +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO comm 0xa48d010 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x782a608dfb3da5ab - Init START +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO comm 0xa5f5f50 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x782a608dfb3da5ab - Init START +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:850787:851326 [2] NCCL INFO comm 0xc44aa20 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8b4b1a338ac2986d - Init START +ip-26-0-163-58:850791:851327 [6] NCCL INFO comm 0xa99dc70 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8b4b1a338ac2986d - Init START +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850791:851327 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:850787:851326 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO comm 0xc3cb5a0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xce4daef07130e6 - Init START +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO comm 0xb699160 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xce4daef07130e6 - Init START +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO comm 0xb69fe60 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1427b4cd904c6e60 - Init START +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO comm 0xb693ae0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1427b4cd904c6e60 - Init START +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO comm 0xa8abd10 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x866b2f4830318ca - Init START +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO comm 0xbbd1de0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x866b2f4830318ca - Init START +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Using network Libfabric +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO comm 0xa6cc0f0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8ad1b12fa0640045 - Init START +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO comm 0xa936b00 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8ad1b12fa0640045 - Init START +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO comm 0xa51dc60 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc584a81ce48bd6a7 - Init START +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO comm 0xa68d830 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc584a81ce48bd6a7 - Init START +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO comm 0xb01e870 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x59e09f47b6e41ad4 - Init START +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO comm 0xb999f60 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x59e09f47b6e41ad4 - Init START +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO comm 0xc065740 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x573a6fba2eaf0d95 - Init START +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO comm 0xc017280 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x573a6fba2eaf0d95 - Init START +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Using network Libfabric +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO comm 0xb5d0340 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xff0dc28e6773902c - Init START +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO comm 0xbea7180 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xff0dc28e6773902c - Init START +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Using network Libfabric +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO comm 0xb1e3d10 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x991e13e8c1b3b9a5 - Init START +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO comm 0xbb02880 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x991e13e8c1b3b9a5 - Init START +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851326 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851327 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-58:850791:851327 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-58:850791:851327 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-58:850787:851326 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850787:851326 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851321 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850788:851323 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:850785:851317 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-58:850792:851324 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850789:851318 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-58:850792:851324 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-58:850792:851324 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-58:850788:851323 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850788:851323 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850789:851318 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-58:850789:851318 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-58:850785:851317 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850785:851317 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850786:851320 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-58:850790:851321 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-58:850790:851321 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-58:850786:851320 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:850786:851320 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851327 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851327 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850787:851326 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:850791:851327 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851327 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851327 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851327 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851327 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850791:851327 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:850792:851324 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851324 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851324 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851324 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851324 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851324 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851324 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850792:851324 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 00/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 01/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850788:851323 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:851318 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:851318 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850785:851317 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:850789:851318 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851318 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851318 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851318 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851318 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850789:851318 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850786:851320 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:850790:851321 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851321 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851321 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851321 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851321 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851321 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851321 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:850790:851321 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Connected all rings +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO Connected all trees +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Connected all rings +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO Connected all trees +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137568:2138088 [6] NCCL INFO comm 0xb699160 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xce4daef07130e6 - Init COMPLETE +ip-26-0-169-139:2137564:2138087 [2] NCCL INFO comm 0xc3cb5a0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xce4daef07130e6 - Init COMPLETE +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Connected all rings +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO Connected all trees +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Connected all rings +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO Connected all trees +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594325:1594843 [4] NCCL INFO comm 0xb802f00 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe8409e998f1197a1 - Init COMPLETE +ip-26-0-164-75:1594321:1594842 [0] NCCL INFO comm 0xc1188e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe8409e998f1197a1 - Init COMPLETE +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Connected all rings +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO Connected all trees +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Connected all rings +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO Connected all trees +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Connected all rings +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO Connected all trees +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Connected all rings +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO Connected all trees +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Connected all rings +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO Connected all trees +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594322:1594846 [1] NCCL INFO comm 0xb01e870 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x59e09f47b6e41ad4 - Init COMPLETE +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Connected all rings +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO Connected all trees +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137566:2138085 [4] NCCL INFO comm 0xb836490 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4beee5f9ade2bcab - Init COMPLETE +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Connected all rings +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO Connected all trees +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594326:1594849 [5] NCCL INFO comm 0xb999f60 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x59e09f47b6e41ad4 - Init COMPLETE +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Connected all rings +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO Connected all trees +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Connected all rings +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO Connected all trees +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137563:2138090 [1] NCCL INFO comm 0xb69fe60 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1427b4cd904c6e60 - Init COMPLETE +ip-26-0-169-139:2137562:2138084 [0] NCCL INFO comm 0xb10ce20 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4beee5f9ade2bcab - Init COMPLETE +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Connected all rings +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO Connected all trees +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Connected all rings +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO Connected all trees +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Connected all rings +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO Connected all trees +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Connected all rings +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO Connected all trees +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Connected all rings +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO Connected all trees +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594327:1594852 [6] NCCL INFO comm 0xb1e3d10 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x991e13e8c1b3b9a5 - Init COMPLETE +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Connected all rings +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO Connected all trees +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-139:2137567:2138091 [5] NCCL INFO comm 0xb693ae0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1427b4cd904c6e60 - Init COMPLETE +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Connected all rings +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO Connected all trees +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594323:1594851 [2] NCCL INFO comm 0xbb02880 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x991e13e8c1b3b9a5 - Init COMPLETE +ip-26-0-169-139:2137569:2138094 [7] NCCL INFO comm 0xa936b00 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8ad1b12fa0640045 - Init COMPLETE +ip-26-0-169-139:2137565:2138093 [3] NCCL INFO comm 0xa6cc0f0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8ad1b12fa0640045 - Init COMPLETE +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Connected all rings +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO Connected all trees +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594324:1594847 [3] NCCL INFO comm 0xa68d830 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc584a81ce48bd6a7 - Init COMPLETE +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Connected all rings +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO Connected all trees +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115365:2115886 [0] NCCL INFO comm 0xa5f5f50 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x782a608dfb3da5ab - Init COMPLETE +ip-26-0-169-207:2115371:2115890 [6] NCCL INFO comm 0xa8abd10 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x866b2f4830318ca - Init COMPLETE +ip-26-0-169-207:2115369:2115887 [4] NCCL INFO comm 0xa48d010 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x782a608dfb3da5ab - Init COMPLETE +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Connected all rings +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO Connected all trees +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115367:2115889 [2] NCCL INFO comm 0xbbd1de0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x866b2f4830318ca - Init COMPLETE +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Connected all rings +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO Connected all trees +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-164-75:1594328:1594848 [7] NCCL INFO comm 0xa51dc60 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc584a81ce48bd6a7 - Init COMPLETE +ip-26-0-169-207:2115370:2115896 [5] NCCL INFO comm 0xb5d0340 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xff0dc28e6773902c - Init COMPLETE +ip-26-0-169-207:2115366:2115895 [1] NCCL INFO comm 0xbea7180 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xff0dc28e6773902c - Init COMPLETE +ip-26-0-163-58:850787:851326 [2] NCCL INFO Connected all rings +ip-26-0-163-58:850787:851326 [2] NCCL INFO Connected all trees +ip-26-0-163-58:850787:851326 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:850787:851326 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-207:2115368:2115892 [3] NCCL INFO comm 0xc065740 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x573a6fba2eaf0d95 - Init COMPLETE +ip-26-0-169-207:2115372:2115893 [7] NCCL INFO comm 0xc017280 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x573a6fba2eaf0d95 - Init COMPLETE +ip-26-0-163-58:850791:851327 [6] NCCL INFO Connected all rings +ip-26-0-163-58:850791:851327 [6] NCCL INFO Connected all trees +ip-26-0-163-58:850791:851327 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:850791:851327 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850787:851326 [2] NCCL INFO comm 0xc44aa20 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8b4b1a338ac2986d - Init COMPLETE +ip-26-0-163-58:850791:851327 [6] NCCL INFO comm 0xa99dc70 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8b4b1a338ac2986d - Init COMPLETE +ip-26-0-163-58:850788:851323 [3] NCCL INFO Connected all rings +ip-26-0-163-58:850788:851323 [3] NCCL INFO Connected all trees +ip-26-0-163-58:850788:851323 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:850788:851323 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850786:851320 [1] NCCL INFO Connected all rings +ip-26-0-163-58:850786:851320 [1] NCCL INFO Connected all trees +ip-26-0-163-58:850786:851320 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:850786:851320 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850792:851324 [7] NCCL INFO Connected all rings +ip-26-0-163-58:850792:851324 [7] NCCL INFO Connected all trees +ip-26-0-163-58:850792:851324 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:850792:851324 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850790:851321 [5] NCCL INFO Connected all rings +ip-26-0-163-58:850790:851321 [5] NCCL INFO Connected all trees +ip-26-0-163-58:850790:851321 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:850790:851321 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850785:851317 [0] NCCL INFO Connected all rings +ip-26-0-163-58:850785:851317 [0] NCCL INFO Connected all trees +ip-26-0-163-58:850785:851317 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:850785:851317 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850789:851318 [4] NCCL INFO Connected all rings +ip-26-0-163-58:850789:851318 [4] NCCL INFO Connected all trees +ip-26-0-163-58:850789:851318 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:850789:851318 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:850788:851323 [3] NCCL INFO comm 0xc2abdd0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x53d3582ba4770597 - Init COMPLETE +ip-26-0-163-58:850786:851320 [1] NCCL INFO comm 0xb892430 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x32efecc46f3158e9 - Init COMPLETE +ip-26-0-163-58:850792:851324 [7] NCCL INFO comm 0xb85f170 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x53d3582ba4770597 - Init COMPLETE +ip-26-0-163-58:850790:851321 [5] NCCL INFO comm 0xb07be60 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x32efecc46f3158e9 - Init COMPLETE +ip-26-0-163-58:850789:851318 [4] NCCL INFO comm 0xb205350 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8863d0fafd732fbd - Init COMPLETE +ip-26-0-163-58:850785:851317 [0] NCCL INFO comm 0xc407f70 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8863d0fafd732fbd - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward +trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + return forward_call(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self.act(gate_states) * up_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/activations.py", line 149, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/activations.py", line 149, in forward + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return nn.functional.silu(input) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/functional.py", line 2072, in silu + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return nn.functional.silu(input) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/functional.py", line 2072, in silu + return torch._C._nn.silu(input) + torch.cuda.outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)OutOfMemoryError +: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 1 has a total capacty of 79.33 GiB of which 1.60 GiB is free. Including non-PyTorch memory, this process has 77.72 GiB memory in use. Of the allocated memory 75.13 GiB is allocated by PyTorch, and 764.65 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return torch._C._nn.silu(input) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 3 has a total capacty of 79.33 GiB of which 1.52 GiB is free. Including non-PyTorch memory, this process has 77.80 GiB memory in use. Of the allocated memory 75.13 GiB is allocated by PyTorch, and 764.65 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self._call_impl(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return self.act(gate_states) * up_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/activations.py", line 149, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return nn.functional.silu(input) +return forward_call(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/functional.py", line 2072, in silu + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return torch._C._nn.silu(input) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 2 has a total capacty of 79.33 GiB of which 1.60 GiB is free. Including non-PyTorch memory, this process has 77.72 GiB memory in use. Of the allocated memory 75.13 GiB is allocated by PyTorch, and 764.65 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + return self.act(gate_states) * up_states File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + + torch.cudareturn self.act(gate_states) * up_states +.hidden_states = self.down_proj(self.split_silu_mul(merged_states))OutOfMemoryError +: torch.cudaCUDA out of memory. Tried to allocate 1.75 GiB. GPU 0 has a total capacty of 79.33 GiB of which 1.94 MiB is free. Including non-PyTorch memory, this process has 79.32 GiB memory in use. Of the allocated memory 76.88 GiB is allocated by PyTorch, and 764.65 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF. +OutOfMemoryError File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 6 has a total capacty of 79.33 GiB of which 97.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 76.88 GiB is allocated by PyTorch, and 508.65 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return self.act(gate_states) * up_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states = self.down_proj(self.split_silu_mul(merged_states))return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/activations.py", line 149, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return nn.functional.silu(input) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/functional.py", line 2072, in silu + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.act(gate_states) * up_states + return torch._C._nn.silu(input)torch.cuda +.OutOfMemoryError: torch.cudaCUDA out of memory. Tried to allocate 1.75 GiB. GPU 7 has a total capacty of 79.33 GiB of which 97.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 76.88 GiB is allocated by PyTorch, and 508.65 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 4 has a total capacty of 79.33 GiB of which 1.52 GiB is free. Including non-PyTorch memory, this process has 77.80 GiB memory in use. Of the allocated memory 75.13 GiB is allocated by PyTorch, and 764.65 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/activations.py", line 149, in forward + return nn.functional.silu(input) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/functional.py", line 2072, in silu + return torch._C._nn.silu(input) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 5 has a total capacty of 79.33 GiB of which 1.60 GiB is free. Including non-PyTorch memory, this process has 77.72 GiB memory in use. Of the allocated memory 75.13 GiB is allocated by PyTorch, and 764.65 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[E TraceUtils.h:35] Store is down while updating #105 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #109 with key NCCL_0_trace_end +[E TraceUtils.h:35] Store is down while updating #109 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #54 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #100 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #95 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #44 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #107 with key NCCL_0_trace_end +[E TraceUtils.h:35] Store is down while updating #80 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #61 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #87 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #50 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #92 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #89 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #2 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #45 with key NCCL_0_trace_end +[E TraceUtils.h:35] Store is down while updating #94 with key NCCL_0_trace_start +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=0|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=0|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=0|ip-26-0-169-207]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=0|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=2|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=0|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=2|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=3|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=0|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=0|ip-26-0-169-139]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=2|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=2|ip-26-0-169-207]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=3|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=0|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=2|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=0|ip-26-0-169-207]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=3|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=1|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=1|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=2|ip-26-0-169-207]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=3|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=3|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=3|ip-26-0-169-207]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=0|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=3|ip-26-0-169-139]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=1|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=1|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=1|ip-26-0-169-139]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=0|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=1|ip-26-0-169-139]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=1|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=0|ip-26-0-169-139]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=2|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=3|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=1|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=3|ip-26-0-169-207]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=3|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=2|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=1|ip-26-0-169-207]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=2|ip-26-0-169-139]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=2|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=3|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=3|ip-26-0-169-139]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=2|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=2|ip-26-0-169-139]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=1|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=1|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=1|ip-26-0-169-207]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=3|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=0|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=2|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=1|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=3|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=1|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=0|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=3|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=0|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=2|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=2|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=1|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=3|ip-26-0-164-75]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=3|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=1|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=0|ip-26-0-164-75]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=0|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=2|ip-26-0-164-75]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=1|ip-26-0-164-75]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=3|ip-26-0-164-75]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=1|ip-26-0-164-75]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=2|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=0|ip-26-0-164-75]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=2|ip-26-0-164-75]: No progress made in communication iteration, attempt 1/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=0|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=0|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=0|ip-26-0-169-207]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=0|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=2|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=2|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=2|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=2|ip-26-0-169-207]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=0|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=2|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=0|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=2|ip-26-0-169-207]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=1|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=0|ip-26-0-169-139]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=3|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=0|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=1|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=1|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=0|ip-26-0-169-207]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=3|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=1|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=1|ip-26-0-169-139]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=1|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=3|ip-26-0-169-139]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=0|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=1|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=1|ip-26-0-169-207]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=3|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=1|ip-26-0-169-139]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=2|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=0|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=3|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=0|ip-26-0-169-139]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=2|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=3|ip-26-0-169-207]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=2|ip-26-0-169-139]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=3|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=3|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=2|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=3|ip-26-0-169-139]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=2|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=2|ip-26-0-169-139]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=1|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=1|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=1|ip-26-0-169-207]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=3|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=3|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=3|ip-26-0-169-207]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=3|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=0|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=2|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=3|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=3|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=1|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=0|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=0|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=1|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=3|ip-26-0-164-75]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=2|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Connection reset by peer. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=3|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=0|ip-26-0-164-75]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=1|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=2|ip-26-0-164-75]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=0|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=3|ip-26-0-164-75]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=1|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=1|ip-26-0-164-75]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=2|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=0|ip-26-0-164-75]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=1|ip-26-0-164-75]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=0|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=0|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=2|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=0|ip-26-0-169-207]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=2|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=0|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=2|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=2|ip-26-0-169-207]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=2|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=0|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=0|ip-26-0-169-139]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=3|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=1|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=2|ip-26-0-169-207]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=0|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=0|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=0|ip-26-0-169-207]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=1|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=3|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=1|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=3|ip-26-0-169-139]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=1|ip-26-0-169-139]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=1|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=0|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=1|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=2|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=1|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=1|ip-26-0-169-139]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=1|ip-26-0-169-207]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=0|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=2|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=0|ip-26-0-169-139]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=2|ip-26-0-169-139]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=3|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=3|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=3|ip-26-0-169-207]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=3|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=3|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=2|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=3|ip-26-0-169-139]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=2|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=2|ip-26-0-169-139]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=0|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=0|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=0|ip-26-0-169-139]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=1|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=3|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=1|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=1|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=3|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=1|ip-26-0-169-139]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=0|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=2|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=1|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=3|ip-26-0-169-139]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=0|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=1|ip-26-0-169-139]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=2|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=0|ip-26-0-169-139]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=2|ip-26-0-169-139]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=3|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=2|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=3|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=3|ip-26-0-169-139]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=2|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=2|ip-26-0-169-139]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=1|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=1|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=1|ip-26-0-169-207]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=3|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=3|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=3|ip-26-0-169-207]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=2|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=0|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=2|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=2|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=2|ip-26-0-169-207]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=0|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=2|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=0|ip-26-0-169-207]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=2|ip-26-0-169-207]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=0|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=0|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=0|ip-26-0-169-207]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=1|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=1|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=1|ip-26-0-169-207]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=3|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=3|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=3|ip-26-0-169-207]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=1|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=1|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=1|ip-26-0-169-207]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=3|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=3|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=3|ip-26-0-169-207]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=2|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=2|ip-26-0-164-75]: No progress made in communication iteration, attempt 2/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=3|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=0|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=3|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=3|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=0|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=0|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=2|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=1|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=3|ip-26-0-164-75]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=0|ip-26-0-164-75]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=3|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=1|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=0|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=2|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=1|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=3|ip-26-0-164-75]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=0|ip-26-0-164-75]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=1|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=2|ip-26-0-164-75]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=1|ip-26-0-164-75]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=2|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=1|ip-26-0-164-75]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=2|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=2|ip-26-0-164-75]: No progress made in communication iteration, attempt 3/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=3|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=0|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=3|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=3|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=0|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=0|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=3|ip-26-0-164-75]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=0|ip-26-0-164-75]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=1|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=3|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=0|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=1|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=3|ip-26-0-164-75]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=0|ip-26-0-164-75]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=1|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=1|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=2|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=1|ip-26-0-164-75]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=1|ip-26-0-164-75]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=2|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=2|ip-26-0-164-75]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=2|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=2|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=2|ip-26-0-164-75]: No progress made in communication iteration, attempt 4/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=2|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=2|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=0|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=2|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=0|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=2|ip-26-0-169-207]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=0|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=2|ip-26-0-169-207]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=2|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=0|ip-26-0-169-207]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=2|ip-26-0-169-207]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=0|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=0|ip-26-0-169-139]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=1|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=0|ip-26-0-169-139]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=0|ip-26-0-169-207]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=2|ip-26-0-169-207]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=0|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=0|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=1|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=1|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=0|ip-26-0-169-207]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=1|ip-26-0-169-139]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=2|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=1|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=0|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=1|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=1|ip-26-0-169-139]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=0|ip-26-0-169-207]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=3|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=2|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=1|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=1|ip-26-0-169-139]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=0|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=2|ip-26-0-169-139]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=3|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=1|ip-26-0-169-207]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=1|ip-26-0-169-207]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=1|ip-26-0-169-139]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=0|ip-26-0-169-139]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=2|ip-26-0-169-139]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [WARNING|DP=0|PP=2|TP=3|ip-26-0-169-139]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=0|ip-26-0-169-139]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=0|PP=2|TP=3|ip-26-0-169-139]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=3|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=2|ip-26-0-169-139]: [RECV META] Rank 2: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=3|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=2|ip-26-0-169-139]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 1: [2] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '1:2', but store->get('1:2') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=3|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=3|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=2|ip-26-0-169-139]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=2|TP=3|ip-26-0-169-139]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=2|ip-26-0-169-139]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=1|PP=2|TP=3|ip-26-0-169-139]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=3|ip-26-0-169-207]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=3|ip-26-0-169-207]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=1|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=1|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=3|TP=1|ip-26-0-169-207]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=3|TP=1|ip-26-0-169-207]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=3|ip-26-0-169-207]: [RECV META] Rank 3: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=3|ip-26-0-169-207]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 2: [3] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '2:3', but store->get('2:3') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=3|TP=3|ip-26-0-169-207]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=3|TP=3|ip-26-0-169-207]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=0|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=0|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=3|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=0|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=0|ip-26-0-164-75]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=3|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=0|ip-26-0-164-75]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=0|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=3|ip-26-0-164-75]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=1|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=0|ip-26-0-164-75]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=3|ip-26-0-164-75]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=0|ip-26-0-164-75]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=2|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=1|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=1|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=1|ip-26-0-164-75]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=2|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=1|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=1|ip-26-0-164-75]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [WARNING|DP=0|PP=1|TP=2|ip-26-0-164-75]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=1|ip-26-0-164-75]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=0|PP=1|TP=2|ip-26-0-164-75]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=1|ip-26-0-164-75]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=2|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=2|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=2|ip-26-0-164-75]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=2|ip-26-0-164-75]: Max retries reached, giving up on communication +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=3|ip-26-0-164-75]: [RECV META] Rank 1: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=3|ip-26-0-164-75]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 0: [1] is setting up NCCL communicator and retrieving ncclUniqueId from [0] via c10d key-value store by key '0:1', but store->get('0:1') got error: Broken pipe. This may indicate a possible application crash on rank 0 or a network set up issue. +12/27/2024 17:45:16 [WARNING|DP=1|PP=1|TP=3|ip-26-0-164-75]: No progress made in communication iteration, attempt 5/5 +12/27/2024 17:45:16 [ERROR|DP=1|PP=1|TP=3|ip-26-0-164-75]: Max retries reached, giving up on communication +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): +Traceback (most recent call last): + return f(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + output = model(**micro_batch)outputs = self.pipeline_engine.train_batch_iter( + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] +trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs)outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + sharded_logits = self.model( +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader)trainer.train(dataloader) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): +Traceback (most recent call last): + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward +trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs)sharded_logits = self.model( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + trainer.train(dataloader) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + trainer.train(dataloader) + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + trainer.train(dataloader) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +return f(*args, **kwargs) +return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return forward_call(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return f(*args, **kwargs) + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + output = model(**micro_batch)output = model(**micro_batch) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter(output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = model(**micro_batch) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs) +sharded_logits = self.model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter(outputs = self.pipeline_engine.train_batch_iter( + +outputs = self.pipeline_engine.train_batch_iter( File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + activation = pipeline_state.activations_buffer.popleft()return self._call_impl(*args, **kwargs) + +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + IndexErrorhidden_encoder_states = encoder_block(**hidden_encoder_states): +pop from an empty deque + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) return forward_call(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs)output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward +output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +output = model(**micro_batch) +output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + output = model(**micro_batch) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + activation = pipeline_state.activations_buffer.popleft() +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +IndexError: pop from an empty deque + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + sharded_logits = self.model( File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +sharded_logits = self.model( + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + return forward_call(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + hidden_encoder_states = encoder_block(**hidden_encoder_states)hidden_encoder_states = encoder_block(**hidden_encoder_states) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + activation = pipeline_state.activations_buffer.popleft() + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +IndexError: pop from an empty deque + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states)activation = pipeline_state.activations_buffer.popleft() + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +IndexError: pop from an empty deque +hidden_encoder_states = encoder_block(**hidden_encoder_states) + hidden_encoder_states = encoder_block(**hidden_encoder_states)hidden_encoder_states = encoder_block(**hidden_encoder_states) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs)return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() + activation = pipeline_state.activations_buffer.popleft()activation = pipeline_state.activations_buffer.popleft() + +IndexErrorIndexError: pop from an empty deque: +pop from an empty deque +IndexError: pop from an empty deque + new_kwargs[name] = recv_from_pipeline_state_buffer(new_kwargs[name] = recv_from_pipeline_state_buffer( + + new_kwargs[name] = recv_from_pipeline_state_buffer( File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + activation = pipeline_state.activations_buffer.popleft() + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer +IndexError: pop from an empty deque + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + activation = pipeline_state.activations_buffer.popleft() +IndexError : activation = pipeline_state.activations_buffer.popleft()pop from an empty deque + +IndexError: pop from an empty deque + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 135, in forward + new_kwargs[name] = recv_from_pipeline_state_buffer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/functional.py", line 124, in recv_from_pipeline_state_buffer + activation = pipeline_state.activations_buffer.popleft() +IndexError: pop from an empty deque +[2024-12-27 17:45:37,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 850786 closing signal SIGTERM +[2024-12-27 17:45:37,813] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 850788 closing signal SIGTERM +[2024-12-27 17:45:37,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 850789 closing signal SIGTERM +[2024-12-27 17:45:37,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 850790 closing signal SIGTERM +[2024-12-27 17:45:37,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 850791 closing signal SIGTERM +[2024-12-27 17:45:37,814] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 850792 closing signal SIGTERM +[2024-12-27 17:45:40,208] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 850785) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 17:45:40,251] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_8dkdr4j7/13803461_6abjg6f_/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:45:14 + host : ip-26-0-163-58.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 850787) + error_file: /tmp/torchelastic_8dkdr4j7/13803461_6abjg6f_/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/activations.py", line 149, in forward + return nn.functional.silu(input) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/functional.py", line 2072, in silu + return torch._C._nn.silu(input) + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 2 has a total capacty of 79.33 GiB of which 1.60 GiB is free. Including non-PyTorch memory, this process has 77.72 GiB memory in use. Of the allocated memory 75.13 GiB is allocated by PyTorch, and 764.65 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:45:14 + host : ip-26-0-163-58.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 850785) + error_file: /tmp/torchelastic_8dkdr4j7/13803461_6abjg6f_/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 257, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.75 GiB. GPU 0 has a total capacty of 79.33 GiB of which 1.94 MiB is free. Including non-PyTorch memory, this process has 79.32 GiB memory in use. Of the allocated memory 76.88 GiB is allocated by PyTorch, and 764.65 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +============================================================ +srun: error: ip-26-0-163-58: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13803461.0 +[2024-12-27 17:45:40,708] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-27 17:45:40,708] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1594322 closing signal SIGTERM +[2024-12-27 17:45:40,708] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1594323 closing signal SIGTERM +[2024-12-27 17:45:40,708] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1594324 closing signal SIGTERM +[2024-12-27 17:45:40,708] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1594325 closing signal SIGTERM +[2024-12-27 17:45:40,708] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1594327 closing signal SIGTERM +[2024-12-27 17:45:40,708] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1594328 closing signal SIGTERM +[2024-12-27 17:45:40,709] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-27 17:45:40,710] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2115369 closing signal SIGTERM +[2024-12-27 17:45:40,709] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-27 17:45:40,710] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2115371 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2137491 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2115293 got signal: 15 +srun: error: ip-26-0-169-139: task 2: Exited with exit code 1 +srun: error: ip-26-0-169-207: task 3: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1594249 got signal: 15 +srun: error: ip-26-0-164-75: task 1: Exited with exit code 1 diff --git a/logs/13803565-bench_3.57G_dp4_tp4_pp2_acc2_mbs32_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13803565-bench_3.57G_dp4_tp4_pp2_acc2_mbs32_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..55b1ed271fca061d52770f639e7e12de23423527 --- /dev/null +++ b/logs/13803565-bench_3.57G_dp4_tp4_pp2_acc2_mbs32_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,10796 @@ ++ '[' -z 13803565 ']' ++ unset FI_PROVIDER ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-170-[31,132,143,160]' ++ export 'NODELIST=ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160' ++ NODELIST='ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-170-[31,132,143,160]' ++ export MASTER_NODE=ip-26-0-170-31 ++ MASTER_NODE=ip-26-0-170-31 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ export NCCL_P2P_LEVEL=LOC ++ NCCL_P2P_LEVEL=LOC ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-170-31' +Master node: ip-26-0-170-31 ++ echo 'All nodes: ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160' +All nodes: ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 ++ echo 'World size: 32' +World size: 32 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13803565 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-170-31:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_3.57G_dp4_tp4_pp2_acc2_mbs32_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-27 17:57:36,018] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:57:36,020] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:57:36,028] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:57:36,020] torch.distributed.run: [WARNING] +[2024-12-27 17:57:36,020] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:57:36,020] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:57:36,020] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:57:36,018] torch.distributed.run: [WARNING] +[2024-12-27 17:57:36,018] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:57:36,018] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:57:36,018] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:57:36,028] torch.distributed.run: [WARNING] +[2024-12-27 17:57:36,028] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:57:36,028] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:57:36,028] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:57:36,154] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 17:57:36,154] torch.distributed.run: [WARNING] +[2024-12-27 17:57:36,154] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 17:57:36,154] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 17:57:36,154] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-170-132:3196831:3196831 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.170.132<0> +ip-26-0-170-132:3196831:3196831 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-132:3196831:3196831 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-132:3196831:3196831 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-132:3196834:3196834 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-132:3196836:3196836 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-132:3196835:3196835 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-132:3196833:3196833 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-132:3196837:3196837 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2134476:2134476 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2134481:2134481 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2134479:2134479 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-132:3196838:3196838 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-132:3196832:3196832 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2098600:2098600 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2098602:2098602 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2134482:2134482 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2134478:2134478 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2098598:2098598 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2134477:2134477 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2134480:2134480 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:2134483:2134483 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-31:3066377:3066377 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-31:3066373:3066373 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-31:3066372:3066372 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-31:3066379:3066379 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-31:3066378:3066378 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-31:3066375:3066375 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-31:3066374:3066374 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-31:3066376:3066376 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2098597:2098597 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2098601:2098601 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-132:3196836:3196836 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.170.132<0> +ip-26-0-170-132:3196834:3196834 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.170.132<0> +ip-26-0-170-132:3196835:3196835 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.170.132<0> +ip-26-0-170-132:3196833:3196833 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.170.132<0> +ip-26-0-170-132:3196837:3196837 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.170.132<0> +ip-26-0-170-132:3196838:3196838 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.170.132<0> +ip-26-0-170-132:3196832:3196832 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.170.132<0> +ip-26-0-170-132:3196833:3196833 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-132:3196837:3196837 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-132:3196836:3196836 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-132:3196834:3196834 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-132:3196835:3196835 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-132:3196833:3196833 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-132:3196837:3196837 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-132:3196834:3196834 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-132:3196836:3196836 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-132:3196835:3196835 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2098596:2098596 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-132:3196838:3196838 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-132:3196838:3196838 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-132:3196832:3196832 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-132:3196832:3196832 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2098602:2098602 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2098599:2098599 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2098600:2098600 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2098598:2098598 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2098603:2098603 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2098602:2098602 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2098602:2098602 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2098600:2098600 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2098600:2098600 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2098598:2098598 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2098598:2098598 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2098597:2098597 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-160:2134476:2134476 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2134479:2134479 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2134481:2134481 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-143:2098601:2098601 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2098597:2098597 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2098597:2098597 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2098601:2098601 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2098601:2098601 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2134476:2134476 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2134476:2134476 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2134482:2134482 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2134478:2134478 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2134479:2134479 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2134479:2134479 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2134481:2134481 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2134481:2134481 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2134477:2134477 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:2134480:2134480 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-143:2098596:2098596 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-160:2134483:2134483 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-143:2098596:2098596 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2098596:2098596 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2134482:2134482 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2134478:2134478 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2134482:2134482 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2134478:2134478 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2098599:2098599 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2098603:2098603 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-160:2134480:2134480 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2134483:2134483 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2134480:2134480 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2134483:2134483 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:2134477:2134477 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:2134477:2134477 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2098603:2098603 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2098603:2098603 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2098599:2098599 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2098599:2098599 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-31:3066374:3066374 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.170.31<0> +ip-26-0-170-31:3066376:3066376 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.170.31<0> +ip-26-0-170-31:3066377:3066377 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.170.31<0> +ip-26-0-170-31:3066373:3066373 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.170.31<0> +ip-26-0-170-31:3066375:3066375 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.170.31<0> +ip-26-0-170-31:3066372:3066372 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.170.31<0> +ip-26-0-170-31:3066379:3066379 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.170.31<0> +ip-26-0-170-31:3066378:3066378 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.170.31<0> +ip-26-0-170-31:3066374:3066374 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-31:3066376:3066376 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-31:3066374:3066374 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-31:3066376:3066376 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-31:3066373:3066373 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-31:3066377:3066377 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-31:3066377:3066377 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-31:3066373:3066373 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-31:3066375:3066375 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-31:3066375:3066375 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-31:3066372:3066372 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-31:3066379:3066379 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-31:3066372:3066372 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-31:3066379:3066379 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-31:3066378:3066378 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-31:3066378:3066378 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/253 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO comm 0x9a38300 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO comm 0x8839740 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO comm 0x8ba8da0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO comm 0x92cdb00 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO comm 0x9b4e400 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO comm 0x9fbf700 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO comm 0x8f48a60 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO comm 0x9104040 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO comm 0x9a41460 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO comm 0x88be6d0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO comm 0x99a7390 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO comm 0x9f3eff0 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO comm 0x8c3f980 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO comm 0x9724230 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO comm 0x8ba2420 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO comm 0x8e62150 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO comm 0x8b47f40 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO comm 0x97e6ed0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO comm 0x8874310 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO comm 0x906c520 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO comm 0x93126c0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO comm 0x9a55740 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO comm 0x865f3c0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO comm 0x9712210 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO comm 0x96b55d0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO comm 0xa094530 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO comm 0x8e50e30 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO comm 0x84ebc80 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO comm 0x9754850 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO comm 0x88ab6e0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO comm 0x8f4ed90 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO comm 0x9d6bdb0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2c427458d6e3e7a6 - Init START +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 15/-1/-1->31->-1 [3] 15/-1/-1->31->-1 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Trees [0] 8/24/-1->16->0 [1] 8/24/-1->16->0 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Trees [0] 16/-1/-1->0->-1 [1] 16/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Connected all rings +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Connected all rings +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Connected all rings +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Connected all rings +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Connected all rings +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Connected all rings +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Connected all rings +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Connected all rings +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Connected all rings +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Connected all rings +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Connected all rings +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Connected all rings +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Connected all rings +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Connected all rings +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Connected all rings +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Connected all rings +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Connected all rings +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Connected all rings +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO Connected all trees +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066379:3066538 [7] NCCL INFO comm 0x9a38300 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO Connected all trees +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196831:3197093 [0] NCCL INFO comm 0x8e62150 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO Connected all trees +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO Connected all trees +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO Connected all trees +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO Connected all trees +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO Connected all trees +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196837:3197098 [6] NCCL INFO comm 0x9d6bdb0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-143:2098598:2098754 [2] NCCL INFO comm 0x865f3c0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-143:2098602:2098757 [6] NCCL INFO comm 0x906c520 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-132:3196833:3197100 [2] NCCL INFO comm 0x8e50e30 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-160:2134478:2134642 [2] NCCL INFO comm 0x9724230 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-160:2134482:2134641 [6] NCCL INFO comm 0x9104040 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-31:3066374:3066534 [2] NCCL INFO comm 0x9b4e400 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-132:3196832:3197096 [1] NCCL INFO comm 0x9712210 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-132:3196836:3197094 [5] NCCL INFO comm 0x9754850 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO Connected all trees +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO Connected all trees +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO Connected all trees +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098601:2098761 [5] NCCL INFO comm 0x8874310 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134477:2134639 [1] NCCL INFO comm 0x8b47f40 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-143:2098597:2098760 [1] NCCL INFO comm 0x96b55d0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-31:3066378:3066535 [6] NCCL INFO comm 0x8839740 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-31:3066373:3066536 [1] NCCL INFO comm 0x99a7390 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-31:3066377:3066539 [5] NCCL INFO comm 0x9fbf700 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-160:2134481:2134638 [5] NCCL INFO comm 0x88be6d0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO Connected all trees +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO Connected all trees +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO Connected all trees +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196838:3197095 [7] NCCL INFO comm 0x8f4ed90 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-143:2098599:2098756 [3] NCCL INFO comm 0x93126c0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-132:3196834:3197097 [3] NCCL INFO comm 0x84ebc80 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO Connected all trees +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066375:3066540 [3] NCCL INFO comm 0x8ba8da0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO Connected all trees +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO Connected all trees +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066376:3066533 [4] NCCL INFO comm 0x92cdb00 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-143:2098603:2098758 [7] NCCL INFO comm 0x8ba2420 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196835:3197099 [4] NCCL INFO comm 0xa094530 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-160:2134479:2134637 [3] NCCL INFO comm 0x9f3eff0 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-143:2098600:2098759 [4] NCCL INFO comm 0x9a55740 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098596:2098755 [0] NCCL INFO comm 0x88ab6e0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-31:3066372:3066537 [0] NCCL INFO comm 0x8f48a60 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-160:2134483:2134640 [7] NCCL INFO comm 0x8c3f980 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-160:2134476:2134635 [0] NCCL INFO comm 0x97e6ed0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-160:2134480:2134636 [4] NCCL INFO comm 0x9a41460 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2c427458d6e3e7a6 - Init COMPLETE +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO comm 0x8de1fe0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO comm 0x9c6ed60 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO comm 0x9505570 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO comm 0x8a724c0 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO comm 0xa1f5ef0 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO comm 0x90a3830 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO comm 0x9949800 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO comm 0x9d85520 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO comm 0x8e782e0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO comm 0x8aad1a0 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO comm 0x8ddb880 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO comm 0x9089ee0 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO comm 0x9bde680 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO comm 0x9181b00 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO comm 0x9187aa0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO comm 0x9fa5aa0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO comm 0x998dd20 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO comm 0x8725920 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO comm 0xa2cdc60 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO comm 0x933bcb0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO comm 0x8af5330 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO comm 0x92a36c0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO comm 0x9c8c560 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO comm 0x9c781e0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO comm 0x9549270 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO comm 0x8ae4b00 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO comm 0x995ce60 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO comm 0xa178300 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO comm 0x8d7eff0 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO comm 0x9a1dbf0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO comm 0x88980e0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO comm 0x98ef300 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x44011444fc71b5b9 - Init START +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Trees [0] 8/24/-1->16->0 [1] 8/24/-1->16->0 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 15/-1/-1->31->-1 [3] 15/-1/-1->31->-1 +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Trees [0] 16/-1/-1->0->-1 [1] 16/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Connected all rings +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Connected all rings +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Connected all rings +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Connected all rings +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Connected all rings +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Connected all rings +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Connected all rings +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Connected all rings +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Connected all rings +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Connected all rings +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Connected all rings +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Connected all rings +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Connected all rings +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Connected all rings +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Connected all rings +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Connected all rings +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Connected all rings +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Connected all rings +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Connected all rings +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO Connected all trees +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO Connected all trees +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066379:3066764 [7] NCCL INFO comm 0x9c6ed60 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-132:3196831:3197323 [0] NCCL INFO comm 0x90a3830 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO Connected all trees +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO Connected all trees +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO Connected all trees +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO Connected all trees +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066374:3066767 [2] NCCL INFO comm 0x9d85520 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO Connected all trees +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196833:3197328 [2] NCCL INFO comm 0x9089ee0 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-132:3196837:3197327 [6] NCCL INFO comm 0x9fa5aa0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-132:3196832:3197330 [1] NCCL INFO comm 0x9949800 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-160:2134478:2134867 [2] NCCL INFO comm 0x995ce60 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-31:3066378:3066762 [6] NCCL INFO comm 0x8a724c0 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO Connected all trees +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134482:2134865 [6] NCCL INFO comm 0x933bcb0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196836:3197326 [5] NCCL INFO comm 0x998dd20 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO Connected all trees +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134481:2134868 [5] NCCL INFO comm 0x8af5330 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-31:3066377:3066763 [5] NCCL INFO comm 0xa1f5ef0 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO Connected all trees +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO Connected all trees +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO Connected all trees +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066373:3066768 [1] NCCL INFO comm 0x9bde680 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO Connected all trees +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066375:3066765 [3] NCCL INFO comm 0x8de1fe0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-132:3196838:3197325 [7] NCCL INFO comm 0x9187aa0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO Connected all trees +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066376:3066766 [4] NCCL INFO comm 0x9505570 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-132:3196834:3197329 [3] NCCL INFO comm 0x8725920 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO Connected all trees +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134477:2134870 [1] NCCL INFO comm 0x8d7eff0 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134479:2134869 [3] NCCL INFO comm 0xa178300 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO Connected all trees +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196835:3197324 [4] NCCL INFO comm 0xa2cdc60 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-160:2134480:2134866 [4] NCCL INFO comm 0x9c781e0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-160:2134483:2134871 [7] NCCL INFO comm 0x8e782e0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-31:3066372:3066769 [0] NCCL INFO comm 0x9181b00 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098598:2098989 [2] NCCL INFO comm 0x88980e0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098602:2098987 [6] NCCL INFO comm 0x92a36c0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098599:2098991 [3] NCCL INFO comm 0x9549270 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-160:2134476:2134864 [0] NCCL INFO comm 0x9a1dbf0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098603:2098986 [7] NCCL INFO comm 0x8ddb880 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098597:2098990 [1] NCCL INFO comm 0x98ef300 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098601:2098985 [5] NCCL INFO comm 0x8aad1a0 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-143:2098600:2098984 [4] NCCL INFO comm 0x9c8c560 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-143:2098596:2098988 [0] NCCL INFO comm 0x8ae4b00 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x44011444fc71b5b9 - Init COMPLETE +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO comm 0x919a760 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1999261c599fa583 - Init START +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO comm 0x9fb8760 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1999261c599fa583 - Init START +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO comm 0x90b64f0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1999261c599fa583 - Init START +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO comm 0x99a09e0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1999261c599fa583 - Init START +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO comm 0x87385e0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1999261c599fa583 - Init START +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO comm 0xa2e0920 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x1999261c599fa583 - Init START +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO comm 0x995d880 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1999261c599fa583 - Init START +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO comm 0x909cba0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1999261c599fa583 - Init START +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO comm 0x8df4ca0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x496f4ac6118cde3c - Init START +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO comm 0x9519ca0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x496f4ac6118cde3c - Init START +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO comm 0x9d99070 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x496f4ac6118cde3c - Init START +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO comm 0x8a85210 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x496f4ac6118cde3c - Init START +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO comm 0x91947c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x496f4ac6118cde3c - Init START +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO comm 0x9bf2900 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x496f4ac6118cde3c - Init START +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO comm 0xa209e20 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x496f4ac6118cde3c - Init START +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO comm 0x9c82d20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x496f4ac6118cde3c - Init START +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO comm 0x9901fc0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xdcc71d9191e795d7 - Init START +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO comm 0x92b7c80 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xdcc71d9191e795d7 - Init START +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO comm 0x8af7d60 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xdcc71d9191e795d7 - Init START +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO comm 0x8dee540 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xdcc71d9191e795d7 - Init START +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO comm 0x8abfe60 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xdcc71d9191e795d7 - Init START +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO comm 0x88aada0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xdcc71d9191e795d7 - Init START +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO comm 0x9ca08a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xdcc71d9191e795d7 - Init START +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO comm 0x955d6c0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xdcc71d9191e795d7 - Init START +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO comm 0x996fb20 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xefcbb099bd8b263c - Init START +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO comm 0x8d93110 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xefcbb099bd8b263c - Init START +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO comm 0xa18afc0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xefcbb099bd8b263c - Init START +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO comm 0x9a32940 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xefcbb099bd8b263c - Init START +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO comm 0x8e8afa0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xefcbb099bd8b263c - Init START +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO comm 0x9c8c510 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xefcbb099bd8b263c - Init START +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO comm 0x8b09440 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xefcbb099bd8b263c - Init START +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO comm 0x934f450 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xefcbb099bd8b263c - Init START +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Connected all rings +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Connected all rings +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Connected all rings +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Connected all rings +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Connected all rings +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Connected all rings +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Connected all rings +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Connected all rings +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Connected all rings +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Connected all rings +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Connected all rings +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Connected all rings +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Connected all rings +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Connected all rings +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Connected all rings +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Connected all rings +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO Connected all trees +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066379:3066811 [7] NCCL INFO comm 0x9c82d20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x496f4ac6118cde3c - Init COMPLETE +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO Connected all trees +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066372:3066809 [0] NCCL INFO comm 0x91947c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x496f4ac6118cde3c - Init COMPLETE +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO Connected all trees +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196831:3197371 [0] NCCL INFO comm 0x90b64f0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1999261c599fa583 - Init COMPLETE +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO Connected all trees +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO Connected all trees +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO Connected all trees +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO Connected all trees +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066373:3066813 [1] NCCL INFO comm 0x9bf2900 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x496f4ac6118cde3c - Init COMPLETE +ip-26-0-170-132:3196838:3197372 [7] NCCL INFO comm 0x919a760 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1999261c599fa583 - Init COMPLETE +ip-26-0-170-31:3066378:3066815 [6] NCCL INFO comm 0x8a85210 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x496f4ac6118cde3c - Init COMPLETE +ip-26-0-170-31:3066374:3066812 [2] NCCL INFO comm 0x9d99070 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x496f4ac6118cde3c - Init COMPLETE +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO Connected all trees +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO Connected all trees +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO Connected all trees +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO Connected all trees +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066375:3066810 [3] NCCL INFO comm 0x8df4ca0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x496f4ac6118cde3c - Init COMPLETE +ip-26-0-170-132:3196832:3197377 [1] NCCL INFO comm 0x995d880 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1999261c599fa583 - Init COMPLETE +ip-26-0-170-31:3066376:3066816 [4] NCCL INFO comm 0x9519ca0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x496f4ac6118cde3c - Init COMPLETE +ip-26-0-170-31:3066377:3066814 [5] NCCL INFO comm 0xa209e20 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x496f4ac6118cde3c - Init COMPLETE +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO Connected all trees +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO Connected all trees +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196837:3197376 [6] NCCL INFO comm 0x9fb8760 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1999261c599fa583 - Init COMPLETE +ip-26-0-170-132:3196833:3197378 [2] NCCL INFO comm 0x909cba0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1999261c599fa583 - Init COMPLETE +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO Connected all trees +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO Connected all trees +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196834:3197374 [3] NCCL INFO comm 0x87385e0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1999261c599fa583 - Init COMPLETE +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO Connected all trees +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196836:3197373 [5] NCCL INFO comm 0x99a09e0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1999261c599fa583 - Init COMPLETE +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197375 [4] NCCL INFO comm 0xa2e0920 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x1999261c599fa583 - Init COMPLETE +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134483:2134900 [7] NCCL INFO comm 0x8e8afa0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xefcbb099bd8b263c - Init COMPLETE +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134476:2134899 [0] NCCL INFO comm 0x9a32940 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xefcbb099bd8b263c - Init COMPLETE +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134477:2134906 [1] NCCL INFO comm 0x8d93110 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xefcbb099bd8b263c - Init COMPLETE +ip-26-0-170-160:2134482:2134901 [6] NCCL INFO comm 0x934f450 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xefcbb099bd8b263c - Init COMPLETE +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098603:2099023 [7] NCCL INFO comm 0x8dee540 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xdcc71d9191e795d7 - Init COMPLETE +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134479:2134902 [3] NCCL INFO comm 0xa18afc0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xefcbb099bd8b263c - Init COMPLETE +ip-26-0-170-160:2134481:2134904 [5] NCCL INFO comm 0x8b09440 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xefcbb099bd8b263c - Init COMPLETE +ip-26-0-170-160:2134478:2134903 [2] NCCL INFO comm 0x996fb20 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xefcbb099bd8b263c - Init COMPLETE +ip-26-0-170-160:2134480:2134905 [4] NCCL INFO comm 0x9c8c510 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xefcbb099bd8b263c - Init COMPLETE +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098596:2099017 [0] NCCL INFO comm 0x8af7d60 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xdcc71d9191e795d7 - Init COMPLETE +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098597:2099022 [1] NCCL INFO comm 0x9901fc0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xdcc71d9191e795d7 - Init COMPLETE +ip-26-0-170-143:2098602:2099021 [6] NCCL INFO comm 0x92b7c80 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xdcc71d9191e795d7 - Init COMPLETE +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098598:2099020 [2] NCCL INFO comm 0x88aada0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xdcc71d9191e795d7 - Init COMPLETE +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098599:2099024 [3] NCCL INFO comm 0x955d6c0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xdcc71d9191e795d7 - Init COMPLETE +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098601:2099018 [5] NCCL INFO comm 0x8abfe60 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xdcc71d9191e795d7 - Init COMPLETE +ip-26-0-170-143:2098600:2099019 [4] NCCL INFO comm 0x9ca08a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xdcc71d9191e795d7 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO comm 0xb6e1e10 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6effcc19e54fb369 - Init START +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO comm 0xada0f80 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6effcc19e54fb369 - Init START +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO comm 0xb3b8040 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6effcc19e54fb369 - Init START +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO comm 0xa59c8f0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6effcc19e54fb369 - Init START +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO comm 0xb909cd0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5e6f4d696f049a18 - Init START +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO comm 0xa1644f0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5e6f4d696f049a18 - Init START +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO comm 0xafcba60 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5e6f4d696f049a18 - Init START +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO comm 0x9d3ff10 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5e6f4d696f049a18 - Init START +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO comm 0xaebabc0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcc86a87eeb23aaa5 - Init START +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO comm 0x9cbb390 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcc86a87eeb23aaa5 - Init START +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO comm 0xb196fa0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcc86a87eeb23aaa5 - Init START +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO comm 0xb442470 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcc86a87eeb23aaa5 - Init START +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO comm 0xaba5ef0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x45d6de48753b3c01 - Init START +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO comm 0xac688f0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x45d6de48753b3c01 - Init START +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO comm 0xb3c2450 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x45d6de48753b3c01 - Init START +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO comm 0xa3cac10 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd2a473c770216ba5 - Init START +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO comm 0xae27cd0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd2a473c770216ba5 - Init START +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO comm 0xaa71f10 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd2a473c770216ba5 - Init START +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO comm 0xba15670 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd2a473c770216ba5 - Init START +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO comm 0xa06ca70 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x45d6de48753b3c01 - Init START +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO comm 0xb0a3ab0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x59be2ededc8994c0 - Init START +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO comm 0xa1f02a0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x59be2ededc8994c0 - Init START +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO comm 0x9ec1f50 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x59be2ededc8994c0 - Init START +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO comm 0xa6ba050 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x59be2ededc8994c0 - Init START +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO comm 0xa961d40 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6ab91d3ad6ce2b78 - Init START +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO comm 0x9efa610 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6ab91d3ad6ce2b78 - Init START +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO comm 0x9cacf60 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6ab91d3ad6ce2b78 - Init START +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO comm 0xad03f50 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6ab91d3ad6ce2b78 - Init START +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO comm 0xa4b5d20 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1f48e23ef9b053f8 - Init START +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO comm 0x9b38020 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1f48e23ef9b053f8 - Init START +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO comm 0xa49e950 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1f48e23ef9b053f8 - Init START +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO comm 0xad5dcd0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1f48e23ef9b053f8 - Init START +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 00/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 01/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 02/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 03/0 : 1[5] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 00/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 01/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 02/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 03/0 : 3[7] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 00/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 01/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 02/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 03/0 : 3[3] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Connected all rings +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Connected all rings +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Connected all rings +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Connected all rings +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Connected all rings +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Connected all rings +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Connected all rings +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Connected all rings +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Connected all rings +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 00/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 01/0 : 0[4] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Connected all rings +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Connected all rings +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Connected all rings +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 02/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 03/0 : 3[7] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 00/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 01/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 02/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 03/0 : 2[6] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Connected all rings +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 00/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 01/0 : 2[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Connected all rings +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Connected all rings +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO Connected all trees +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196834:3197416 [3] NCCL INFO comm 0x9b38020 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1f48e23ef9b053f8 - Init COMPLETE +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO Connected all trees +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196831:3197414 [0] NCCL INFO comm 0xa4b5d20 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1f48e23ef9b053f8 - Init COMPLETE +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO Connected all trees +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO Connected all trees +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196838:3197411 [7] NCCL INFO comm 0xa59c8f0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6effcc19e54fb369 - Init COMPLETE +ip-26-0-170-160:2134479:2134944 [3] NCCL INFO comm 0xb3c2450 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x45d6de48753b3c01 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196835:3197409 [4] NCCL INFO comm 0xb6e1e10 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6effcc19e54fb369 - Init COMPLETE +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO comm 0xbe07d30 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf1afd210c4c43c28 - Init START +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO comm 0xa57d8d0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf1afd210c4c43c28 - Init START +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098599:2099063 [3] NCCL INFO comm 0xa961d40 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6ab91d3ad6ce2b78 - Init COMPLETE +ip-26-0-170-143:2098603:2099065 [7] NCCL INFO comm 0xa1f02a0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x59be2ededc8994c0 - Init COMPLETE +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098596:2099059 [0] NCCL INFO comm 0x9efa610 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6ab91d3ad6ce2b78 - Init COMPLETE +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO Connected all trees +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO Connected all trees +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066375:3066855 [3] NCCL INFO comm 0xaa71f10 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd2a473c770216ba5 - Init COMPLETE +ip-26-0-170-143:2098600:2099058 [4] NCCL INFO comm 0xb0a3ab0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x59be2ededc8994c0 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO comm 0xaa8cf80 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x607280c714326e98 - Init START +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO comm 0xb3a7450 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x607280c714326e98 - Init START +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066379:3066853 [7] NCCL INFO comm 0xaebabc0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcc86a87eeb23aaa5 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO Connected all trees +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO comm 0xb900130 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5ff942db577ed539 - Init START +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO comm 0xac35ab0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5ff942db577ed539 - Init START +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134480:2134937 [4] NCCL INFO comm 0xb909cd0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5e6f4d696f049a18 - Init COMPLETE +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO comm 0xb924f10 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7ab830e4a4295742 - Init START +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO comm 0xc127330 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7ab830e4a4295742 - Init START +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134483:2134941 [7] NCCL INFO comm 0xa1644f0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5e6f4d696f049a18 - Init COMPLETE +ip-26-0-170-132:3196833:3197415 [2] NCCL INFO comm 0xa49e950 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1f48e23ef9b053f8 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO comm 0xaba9430 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x958aa598ee09bf04 - Init START +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO comm 0xafe2110 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x958aa598ee09bf04 - Init START +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO Connected all trees +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO Connected all trees +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-132:3196832:3197417 [1] NCCL INFO comm 0xad5dcd0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1f48e23ef9b053f8 - Init COMPLETE +ip-26-0-170-31:3066372:3066850 [0] NCCL INFO comm 0xa3cac10 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd2a473c770216ba5 - Init COMPLETE +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO comm 0xae104b0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf63d20e0ebe85ab7 - Init START +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO comm 0xa940080 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf63d20e0ebe85ab7 - Init START +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098598:2099060 [2] NCCL INFO comm 0x9cacf60 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6ab91d3ad6ce2b78 - Init COMPLETE +ip-26-0-170-143:2098602:2099064 [6] NCCL INFO comm 0xa6ba050 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x59be2ededc8994c0 - Init COMPLETE +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098601:2099061 [5] NCCL INFO comm 0x9ec1f50 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x59be2ededc8994c0 - Init COMPLETE +ip-26-0-170-143:2098597:2099062 [1] NCCL INFO comm 0xad03f50 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6ab91d3ad6ce2b78 - Init COMPLETE +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134476:2134939 [0] NCCL INFO comm 0xac688f0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x45d6de48753b3c01 - Init COMPLETE +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO Connected all trees +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO Connected all trees +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO comm 0xb6ae310 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfc51463187ce6098 - Init START +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO comm 0xaefbc00 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfc51463187ce6098 - Init START +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196837:3197412 [6] NCCL INFO comm 0xb3b8040 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6effcc19e54fb369 - Init COMPLETE +ip-26-0-170-132:3196836:3197413 [5] NCCL INFO comm 0xada0f80 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6effcc19e54fb369 - Init COMPLETE +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO Connected all trees +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066376:3066849 [4] NCCL INFO comm 0xb196fa0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xcc86a87eeb23aaa5 - Init COMPLETE +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO comm 0xb1b1fd0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7dfae65fa3f4e50e - Init START +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO comm 0xbae91d0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7dfae65fa3f4e50e - Init START +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO Connected all trees +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066378:3066852 [6] NCCL INFO comm 0x9cbb390 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcc86a87eeb23aaa5 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO comm 0xa700b20 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1d5335a881fd149b - Init START +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO comm 0xb0ffa20 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1d5335a881fd149b - Init START +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134482:2134942 [6] NCCL INFO comm 0xafcba60 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5e6f4d696f049a18 - Init COMPLETE +ip-26-0-170-160:2134478:2134945 [2] NCCL INFO comm 0xaba5ef0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x45d6de48753b3c01 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO comm 0xafe59b0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfae2972d18ff7a2e - Init START +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO comm 0xbdfd8c0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfae2972d18ff7a2e - Init START +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO comm 0xb5ebb60 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf1ead1bf66234aae - Init START +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO comm 0xaee3fa0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf1ead1bf66234aae - Init START +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134477:2134943 [1] NCCL INFO comm 0xa06ca70 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x45d6de48753b3c01 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO comm 0xaab1a40 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf4734733b986546d - Init START +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO comm 0xb7a36b0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf4734733b986546d - Init START +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO Connected all trees +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066374:3066854 [2] NCCL INFO comm 0xba15670 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd2a473c770216ba5 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO comm 0xba2fd60 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x69fad0cc7ddbc17a - Init START +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO comm 0xa6f2a40 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x69fad0cc7ddbc17a - Init START +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134481:2134940 [5] NCCL INFO comm 0x9d3ff10 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5e6f4d696f049a18 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO comm 0xa785810 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3de667cfbbc515de - Init START +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO comm 0xb7e6a00 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3de667cfbbc515de - Init START +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO Connected all trees +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO Connected all trees +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-31:3066373:3066856 [1] NCCL INFO comm 0xae27cd0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd2a473c770216ba5 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO comm 0xb86d700 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x586363441cd4a19c - Init START +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO comm 0xb7495f0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x586363441cd4a19c - Init START +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066377:3066851 [5] NCCL INFO comm 0xb442470 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcc86a87eeb23aaa5 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO comm 0xbe87ce0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb968b41439bd1d43 - Init START +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO comm 0xa907800 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb968b41439bd1d43 - Init START +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Connected all rings +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO Connected all trees +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Connected all rings +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO Connected all trees +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Connected all rings +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO Connected all trees +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Connected all rings +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO Connected all trees +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196838:3197443 [7] NCCL INFO comm 0xafe2110 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x958aa598ee09bf04 - Init COMPLETE +ip-26-0-170-132:3196834:3197439 [3] NCCL INFO comm 0xa57d8d0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf1afd210c4c43c28 - Init COMPLETE +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134479:2134966 [3] NCCL INFO comm 0xbe07d30 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf1afd210c4c43c28 - Init COMPLETE +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196835:3197441 [4] NCCL INFO comm 0xc127330 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7ab830e4a4295742 - Init COMPLETE +ip-26-0-170-160:2134480:2134967 [4] NCCL INFO comm 0xb924f10 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7ab830e4a4295742 - Init COMPLETE +ip-26-0-170-132:3196831:3197447 [0] NCCL INFO comm 0xaefbc00 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfc51463187ce6098 - Init COMPLETE +ip-26-0-170-160:2134476:2134971 [0] NCCL INFO comm 0xb6ae310 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfc51463187ce6098 - Init COMPLETE +ip-26-0-170-160:2134483:2134968 [7] NCCL INFO comm 0xaba9430 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x958aa598ee09bf04 - Init COMPLETE +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Connected all rings +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO Connected all trees +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098603:2099089 [7] NCCL INFO comm 0xac35ab0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5ff942db577ed539 - Init COMPLETE +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Connected all rings +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO Connected all trees +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196832:3197459 [1] NCCL INFO comm 0xb7a36b0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf4734733b986546d - Init COMPLETE +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134482:2134974 [6] NCCL INFO comm 0xafe59b0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfae2972d18ff7a2e - Init COMPLETE +ip-26-0-170-132:3196837:3197453 [6] NCCL INFO comm 0xbdfd8c0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfae2972d18ff7a2e - Init COMPLETE +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Connected all rings +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO Connected all trees +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098600:2099099 [4] NCCL INFO comm 0xbae91d0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7dfae65fa3f4e50e - Init COMPLETE +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196833:3197455 [2] NCCL INFO comm 0xaee3fa0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf1ead1bf66234aae - Init COMPLETE +ip-26-0-170-160:2134477:2134977 [1] NCCL INFO comm 0xaab1a40 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf4734733b986546d - Init COMPLETE +ip-26-0-170-160:2134478:2134975 [2] NCCL INFO comm 0xb5ebb60 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf1ead1bf66234aae - Init COMPLETE +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Connected all rings +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO Connected all trees +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196836:3197462 [5] NCCL INFO comm 0xb7e6a00 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3de667cfbbc515de - Init COMPLETE +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134481:2134981 [5] NCCL INFO comm 0xa785810 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3de667cfbbc515de - Init COMPLETE +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098599:2099087 [3] NCCL INFO comm 0xb3a7450 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x607280c714326e98 - Init COMPLETE +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098596:2099091 [0] NCCL INFO comm 0xa940080 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf63d20e0ebe85ab7 - Init COMPLETE +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098602:2099101 [6] NCCL INFO comm 0xb0ffa20 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1d5335a881fd149b - Init COMPLETE +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Connected all rings +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO Connected all trees +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Connected all rings +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO Connected all trees +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Connected all rings +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO Connected all trees +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066376:3066884 [4] NCCL INFO comm 0xb1b1fd0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7dfae65fa3f4e50e - Init COMPLETE +ip-26-0-170-31:3066375:3066877 [3] NCCL INFO comm 0xaa8cf80 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x607280c714326e98 - Init COMPLETE +ip-26-0-170-31:3066379:3066878 [7] NCCL INFO comm 0xb900130 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5ff942db577ed539 - Init COMPLETE +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Connected all rings +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO Connected all trees +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Connected all rings +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO Connected all trees +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066372:3066879 [0] NCCL INFO comm 0xae104b0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf63d20e0ebe85ab7 - Init COMPLETE +ip-26-0-170-31:3066378:3066885 [6] NCCL INFO comm 0xa700b20 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1d5335a881fd149b - Init COMPLETE +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Connected all rings +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO Connected all trees +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066374:3066886 [2] NCCL INFO comm 0xba2fd60 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x69fad0cc7ddbc17a - Init COMPLETE +ip-26-0-170-143:2098598:2099103 [2] NCCL INFO comm 0xa6f2a40 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x69fad0cc7ddbc17a - Init COMPLETE +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Connected all rings +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO Connected all trees +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Connected all rings +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO Connected all trees +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066377:3066890 [5] NCCL INFO comm 0xbe87ce0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb968b41439bd1d43 - Init COMPLETE +ip-26-0-170-143:2098601:2099107 [5] NCCL INFO comm 0xa907800 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb968b41439bd1d43 - Init COMPLETE +ip-26-0-170-143:2098597:2099105 [1] NCCL INFO comm 0xb7495f0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x586363441cd4a19c - Init COMPLETE +ip-26-0-170-31:3066373:3066888 [1] NCCL INFO comm 0xb86d700 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x586363441cd4a19c - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO comm 0xa8edad0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb6c32960b4c8d349 - Init START +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO comm 0xbc022b0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb6c32960b4c8d349 - Init START +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO comm 0xb1b7f30 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb6c32960b4c8d349 - Init START +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO comm 0xb7d8300 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb6c32960b4c8d349 - Init START +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO comm 0xb129af0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcb168c06384cc18a - Init START +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO comm 0xbe27be0 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcb168c06384cc18a - Init START +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO comm 0xaf0ef30 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xcb168c06384cc18a - Init START +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO comm 0xa71c710 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xcb168c06384cc18a - Init START +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO comm 0xbb13e80 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x20dde84f7c0e67c9 - Init START +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO comm 0xc151de0 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x20dde84f7c0e67c9 - Init START +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO comm 0xaf2cbc0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x20dde84f7c0e67c9 - Init START +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO comm 0xa96aed0 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x20dde84f7c0e67c9 - Init START +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO comm 0xb383ce0 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb04439ee9ac8122e - Init START +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO comm 0xaffd610 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb04439ee9ac8122e - Init START +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO comm 0xb89b200 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb04439ee9ac8122e - Init START +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO comm 0xbaf6f50 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb04439ee9ac8122e - Init START +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO comm 0xbaed130 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbc7dda97a04ee49c - Init START +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO comm 0xac5eba0 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbc7dda97a04ee49c - Init START +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO comm 0xad95af0 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbc7dda97a04ee49c - Init START +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO comm 0xbff4cf0 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbc7dda97a04ee49c - Init START +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO comm 0xc074e70 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd6d58488002f351d - Init START +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO comm 0xba5a0f0 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd6d58488002f351d - Init START +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO comm 0xa972230 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd6d58488002f351d - Init START +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO comm 0xac9ddd0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd6d58488002f351d - Init START +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO comm 0xa5a7d50 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8d1a3a1c5f6cf64c - Init START +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO comm 0xb00cdc0 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8d1a3a1c5f6cf64c - Init START +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO comm 0xac5fc90 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8d1a3a1c5f6cf64c - Init START +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO comm 0xb3d1f50 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8d1a3a1c5f6cf64c - Init START +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO comm 0xa931b20 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf519b00b76bcf5f - Init START +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO comm 0xb810a40 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf519b00b76bcf5f - Init START +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO comm 0xb7ce120 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf519b00b76bcf5f - Init START +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO comm 0xb773cf0 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf519b00b76bcf5f - Init START +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 00/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 01/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 02/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 03/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 00/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 01/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 02/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 03/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 00/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 01/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 02/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 03/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 00/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 01/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 02/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 03/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 00/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 01/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 02/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 03/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 00/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 01/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 02/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 03/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 00/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 00/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 01/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 02/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 01/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 03/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 00/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 02/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 03/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 01/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 02/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 03/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 00/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 01/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 02/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 03/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 00/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 01/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 02/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 03/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 00/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 01/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 02/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 03/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 00/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 01/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 02/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 03/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 00/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 01/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 02/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 03/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 00/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 00/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 01/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 01/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 02/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 02/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 03/0 : 3[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 03/0 : 2[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 00/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 01/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 02/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 00/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 03/0 : 3[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 01/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 02/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 03/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 00/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 01/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 02/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 03/0 : 2[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 00/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 01/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 02/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 03/0 : 0[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 00/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 01/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 00/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 02/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 01/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 03/0 : 3[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 02/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 03/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 02/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 03/0 : 0[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 00/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 01/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 02/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 03/0 : 2[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 00/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 01/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 02/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 03/0 : 3[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 00/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 01/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 02/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 03/0 : 0[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 00/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 01/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 02/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 03/0 : 2[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 00/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 01/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 02/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 03/0 : 3[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 00/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 01/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 02/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 03/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Connected all rings +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Connected all rings +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 00/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 01/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 02/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 03/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 00/0 : 3[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 01/0 : 3[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Connected all rings +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 00/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 01/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 02/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 03/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Connected all rings +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Connected all rings +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 00/0 : 3[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 01/0 : 3[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Connected all rings +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Connected all rings +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Connected all rings +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 00/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 01/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 02/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 03/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 00/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Connected all rings +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 01/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 02/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 03/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 00/0 : 3[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 01/0 : 3[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 00/0 : 3[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 01/0 : 3[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Connected all rings +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 00/0 : 3[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 01/0 : 3[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Connected all rings +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 00/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 01/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 02/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 03/0 : 2[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Connected all rings +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 00/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 01/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 02/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 03/0 : 2[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Connected all rings +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 02/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 00/0 : 3[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Channel 03/0 : 1[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 01/0 : 3[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Connected all rings +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Connected all rings +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 00/0 : 3[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 01/0 : 3[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Connected all rings +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 00/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 01/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 02/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 03/0 : 2[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 00/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 01/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 02/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 03/0 : 2[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Connected all rings +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 02/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 00/0 : 3[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Channel 03/0 : 1[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 01/0 : 3[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO Connected all trees +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134477:2135002 [1] NCCL INFO comm 0xac9ddd0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd6d58488002f351d - Init COMPLETE +ip-26-0-170-31:3066377:3066910 [5] NCCL INFO comm 0xc074e70 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd6d58488002f351d - Init COMPLETE +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO Connected all trees +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO Connected all trees +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066376:3066907 [4] NCCL INFO comm 0xb383ce0 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb04439ee9ac8122e - Init COMPLETE +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO Connected all trees +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066378:3066905 [6] NCCL INFO comm 0xa8edad0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb6c32960b4c8d349 - Init COMPLETE +ip-26-0-170-160:2134479:2134999 [3] NCCL INFO comm 0xbff4cf0 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbc7dda97a04ee49c - Init COMPLETE +ip-26-0-170-160:2134478:2134993 [2] NCCL INFO comm 0xb7d8300 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb6c32960b4c8d349 - Init COMPLETE +ip-26-0-170-31:3066379:3066908 [7] NCCL INFO comm 0xbaed130 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbc7dda97a04ee49c - Init COMPLETE +ip-26-0-170-160:2134476:2134996 [0] NCCL INFO comm 0xb89b200 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb04439ee9ac8122e - Init COMPLETE +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO Connected all trees +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO Connected all trees +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066373:3066911 [1] NCCL INFO comm 0xba5a0f0 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd6d58488002f351d - Init COMPLETE +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO Connected all trees +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO Connected all trees +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196833:3197472 [2] NCCL INFO comm 0xaf0ef30 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xcb168c06384cc18a - Init COMPLETE +ip-26-0-170-143:2098602:2099121 [6] NCCL INFO comm 0xb129af0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcb168c06384cc18a - Init COMPLETE +ip-26-0-170-31:3066374:3066904 [2] NCCL INFO comm 0xbc022b0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb6c32960b4c8d349 - Init COMPLETE +ip-26-0-170-160:2134481:2135003 [5] NCCL INFO comm 0xa972230 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd6d58488002f351d - Init COMPLETE +ip-26-0-170-31:3066375:3066909 [3] NCCL INFO comm 0xac5eba0 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbc7dda97a04ee49c - Init COMPLETE +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO Connected all trees +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO Connected all trees +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098601:2099126 [5] NCCL INFO comm 0xa931b20 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf519b00b76bcf5f - Init COMPLETE +ip-26-0-170-132:3196832:3197482 [1] NCCL INFO comm 0xb7ce120 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf519b00b76bcf5f - Init COMPLETE +ip-26-0-170-160:2134482:2134994 [6] NCCL INFO comm 0xb1b7f30 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb6c32960b4c8d349 - Init COMPLETE +ip-26-0-170-143:2098603:2099125 [7] NCCL INFO comm 0xac5fc90 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8d1a3a1c5f6cf64c - Init COMPLETE +ip-26-0-170-143:2098600:2099123 [4] NCCL INFO comm 0xbb13e80 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x20dde84f7c0e67c9 - Init COMPLETE +ip-26-0-170-31:3066372:3066906 [0] NCCL INFO comm 0xaffd610 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb04439ee9ac8122e - Init COMPLETE +ip-26-0-170-160:2134483:2135000 [7] NCCL INFO comm 0xad95af0 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbc7dda97a04ee49c - Init COMPLETE +ip-26-0-170-160:2134480:2134997 [4] NCCL INFO comm 0xbaf6f50 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb04439ee9ac8122e - Init COMPLETE +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO Connected all trees +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196831:3197475 [0] NCCL INFO comm 0xaf2cbc0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x20dde84f7c0e67c9 - Init COMPLETE +ip-26-0-170-143:2098598:2099120 [2] NCCL INFO comm 0xa71c710 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xcb168c06384cc18a - Init COMPLETE +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO Connected all trees +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO Connected all trees +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196837:3197473 [6] NCCL INFO comm 0xbe27be0 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xcb168c06384cc18a - Init COMPLETE +ip-26-0-170-143:2098597:2099127 [1] NCCL INFO comm 0xb773cf0 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf519b00b76bcf5f - Init COMPLETE +ip-26-0-170-143:2098599:2099124 [3] NCCL INFO comm 0xb3d1f50 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8d1a3a1c5f6cf64c - Init COMPLETE +ip-26-0-170-132:3196834:3197478 [3] NCCL INFO comm 0xa5a7d50 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8d1a3a1c5f6cf64c - Init COMPLETE +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO Connected all trees +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO Connected all trees +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196836:3197483 [5] NCCL INFO comm 0xb810a40 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf519b00b76bcf5f - Init COMPLETE +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO Connected all trees +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098596:2099122 [0] NCCL INFO comm 0xa96aed0 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x20dde84f7c0e67c9 - Init COMPLETE +ip-26-0-170-132:3196838:3197479 [7] NCCL INFO comm 0xb00cdc0 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x8d1a3a1c5f6cf64c - Init COMPLETE +ip-26-0-170-132:3196835:3197476 [4] NCCL INFO comm 0xc151de0 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x20dde84f7c0e67c9 - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO comm 0xcb7cc80 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe8a4a9e2dd502bb8 - Init START +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO comm 0xc5f4b20 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6a8c2a23852a44f2 - Init START +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO comm 0x1b576cd0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe8a4a9e2dd502bb8 - Init START +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO comm 0x1b9175b0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6a8c2a23852a44f2 - Init START +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO comm 0xb3f6f90 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x99b67a56800d63af - Init START +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO comm 0xbe8c760 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7bf43328eeaa7d9a - Init START +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO comm 0x1d385c10 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x99b67a56800d63af - Init START +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO comm 0x1c6127b0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7bf43328eeaa7d9a - Init START +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO comm 0xc7091e0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc9f28696d98eb171 - Init START +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO comm 0x1b3b5530 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc9f28696d98eb171 - Init START +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO comm 0xbb05e00 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe1a47bf72c74dfec - Init START +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO comm 0x1b621560 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe1a47bf72c74dfec - Init START +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO comm 0xc5669e0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8518e5ab08fb528c - Init START +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO comm 0x1dbf37e0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8518e5ab08fb528c - Init START +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Using network Libfabric +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO comm 0xb768030 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9705c5601a1013ad - Init START +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO comm 0x1c099620 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9705c5601a1013ad - Init START +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO comm 0x1d18d120 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xec5bc96f10d27bca - Init START +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO comm 0xbcbe160 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xec5bc96f10d27bca - Init START +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO comm 0xc5fdf40 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x70de42db2b11d908 - Init START +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO comm 0x1cf10140 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x70de42db2b11d908 - Init START +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO comm 0xb47b0a0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x79ef664d7008eb39 - Init START +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO comm 0x1c213bb0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x79ef664d7008eb39 - Init START +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO comm 0xb89c5e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x767d548c2e19502a - Init START +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO comm 0x1c936c70 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x767d548c2e19502a - Init START +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Using network Libfabric +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO comm 0xc2e1270 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3dfddebeabc86630 - Init START +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO comm 0x1bb519c0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3dfddebeabc86630 - Init START +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO comm 0xc3c0260 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94f911421101c9f1 - Init START +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO comm 0x1c3bc990 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94f911421101c9f1 - Init START +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO comm 0xb7a5070 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x59f4256e9417a433 - Init START +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO comm 0x1bdfbc90 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x59f4256e9417a433 - Init START +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO comm 0xcafb4c0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe5e715d0f488e579 - Init START +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO comm 0x1cb557b0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe5e715d0f488e579 - Init START +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Connected all rings +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO Connected all trees +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196838:3197531 [7] NCCL INFO comm 0x1c936c70 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x767d548c2e19502a - Init COMPLETE +ip-26-0-170-132:3196838:3197557 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-132:3196838:3197557 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Connected all rings +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO Connected all trees +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Connected all rings +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO Connected all trees +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196833:3197538 [2] NCCL INFO comm 0x1bb519c0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3dfddebeabc86630 - Init COMPLETE +ip-26-0-170-132:3196833:3197558 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-132:3196833:3197558 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-160:2134483:2135042 [7] NCCL INFO comm 0xb89c5e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x767d548c2e19502a - Init COMPLETE +ip-26-0-170-160:2134483:2135064 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-160:2134483:2135064 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Connected all rings +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO Connected all trees +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Connected all rings +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO Connected all trees +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Connected all rings +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO Connected all trees +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Connected all rings +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO Connected all trees +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Connected all rings +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO Connected all trees +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Connected all rings +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO Connected all trees +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Connected all rings +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO Connected all trees +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134478:2135043 [2] NCCL INFO comm 0xc2e1270 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3dfddebeabc86630 - Init COMPLETE +ip-26-0-170-160:2134478:2135065 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-160:2134478:2135065 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-31:3066378:3067041 [6] NCCL INFO comm 0xb3f6f90 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x99b67a56800d63af - Init COMPLETE +ip-26-0-170-31:3066378:3067064 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-31:3066378:3067064 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-132:3196831:3197537 [0] NCCL INFO comm 0x1c3bc990 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94f911421101c9f1 - Init COMPLETE +ip-26-0-170-31:3066373:3067044 [1] NCCL INFO comm 0xc5669e0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8518e5ab08fb528c - Init COMPLETE +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Connected all rings +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO Connected all trees +ip-26-0-170-132:3196831:3197559 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196831:3197559 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-31:3066373:3067065 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-31:3066373:3067065 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-31:3066375:3067046 [3] NCCL INFO comm 0xb768030 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9705c5601a1013ad - Init COMPLETE +ip-26-0-170-31:3066372:3067045 [0] NCCL INFO comm 0xbb05e00 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe1a47bf72c74dfec - Init COMPLETE +ip-26-0-170-31:3066375:3067066 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-31:3066375:3067066 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-31:3066374:3067043 [2] NCCL INFO comm 0xc7091e0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc9f28696d98eb171 - Init COMPLETE +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Connected all rings +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO Connected all trees +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066372:3067067 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-31:3066372:3067067 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-31:3066374:3067068 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-31:3066374:3067068 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Connected all rings +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO Connected all trees +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Connected all rings +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO Connected all trees +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Connected all rings +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO Connected all trees +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Connected all rings +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO Connected all trees +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Connected all rings +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO Connected all trees +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Connected all rings +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO Connected all trees +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066376:3067039 [4] NCCL INFO comm 0xbe8c760 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7bf43328eeaa7d9a - Init COMPLETE +ip-26-0-170-31:3066376:3067069 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-132:3196836:3197530 [5] NCCL INFO comm 0x1c213bb0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x79ef664d7008eb39 - Init COMPLETE +ip-26-0-170-31:3066376:3067069 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-132:3196834:3197539 [3] NCCL INFO comm 0x1cb557b0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe5e715d0f488e579 - Init COMPLETE +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Connected all rings +ip-26-0-170-132:3196836:3197560 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO Connected all trees +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196836:3197560 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-132:3196834:3197561 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-132:3196834:3197561 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196832:3197536 [1] NCCL INFO comm 0x1bdfbc90 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x59f4256e9417a433 - Init COMPLETE +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Connected all rings +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO Connected all trees +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196832:3197562 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-132:3196832:3197562 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-160:2134481:2135039 [5] NCCL INFO comm 0xb47b0a0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x79ef664d7008eb39 - Init COMPLETE +ip-26-0-170-143:2098602:2099173 [6] NCCL INFO comm 0x1d385c10 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x99b67a56800d63af - Init COMPLETE +ip-26-0-170-160:2134481:2135066 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-160:2134481:2135066 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-143:2098602:2099201 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2098602:2099201 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Connected all rings +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO Connected all trees +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Connected all rings +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO Connected all trees +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Connected all rings +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO Connected all trees +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Connected all rings +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO Connected all trees +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134476:2135044 [0] NCCL INFO comm 0xc3c0260 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94f911421101c9f1 - Init COMPLETE +ip-26-0-170-160:2134476:2135067 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-31:3066379:3067042 [7] NCCL INFO comm 0xc5f4b20 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6a8c2a23852a44f2 - Init COMPLETE +ip-26-0-170-160:2134476:2135067 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-31:3066379:3067070 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-31:3066379:3067070 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2098598:2099178 [2] NCCL INFO comm 0x1b3b5530 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc9f28696d98eb171 - Init COMPLETE +ip-26-0-170-143:2098598:2099202 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-143:2098598:2099202 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-170-132:3196837:3197528 [6] NCCL INFO comm 0x1d18d120 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xec5bc96f10d27bca - Init COMPLETE +ip-26-0-170-132:3196837:3197563 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-132:3196837:3197563 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2098597:2099182 [1] NCCL INFO comm 0x1dbf37e0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8518e5ab08fb528c - Init COMPLETE +ip-26-0-170-143:2098597:2099203 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-143:2098597:2099203 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-132:3196835:3197529 [4] NCCL INFO comm 0x1cf10140 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x70de42db2b11d908 - Init COMPLETE +ip-26-0-170-132:3196835:3197564 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2098599:2099180 [3] NCCL INFO comm 0x1c099620 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9705c5601a1013ad - Init COMPLETE +ip-26-0-170-132:3196835:3197564 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2098599:2099204 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2098599:2099204 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2098596:2099181 [0] NCCL INFO comm 0x1b621560 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe1a47bf72c74dfec - Init COMPLETE +ip-26-0-170-160:2134479:2135046 [3] NCCL INFO comm 0xcafb4c0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe5e715d0f488e579 - Init COMPLETE +ip-26-0-170-160:2134479:2135068 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2098596:2099205 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-160:2134479:2135068 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-170-143:2098596:2099205 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:2134480:2135040 [4] NCCL INFO comm 0xc5fdf40 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x70de42db2b11d908 - Init COMPLETE +ip-26-0-170-160:2134480:2135069 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-160:2134477:2135045 [1] NCCL INFO comm 0xb7a5070 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x59f4256e9417a433 - Init COMPLETE +ip-26-0-170-160:2134480:2135069 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-31:3066377:3067040 [5] NCCL INFO comm 0xcb7cc80 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe8a4a9e2dd502bb8 - Init COMPLETE +ip-26-0-170-160:2134482:2135041 [6] NCCL INFO comm 0xbcbe160 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xec5bc96f10d27bca - Init COMPLETE +ip-26-0-170-160:2134477:2135070 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-160:2134477:2135070 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-31:3066377:3067071 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-31:3066377:3067071 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-160:2134482:2135071 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-160:2134482:2135071 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2098600:2099171 [4] NCCL INFO comm 0x1c6127b0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7bf43328eeaa7d9a - Init COMPLETE +ip-26-0-170-143:2098600:2099206 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2098600:2099206 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-170-143:2098603:2099174 [7] NCCL INFO comm 0x1b9175b0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6a8c2a23852a44f2 - Init COMPLETE +ip-26-0-170-143:2098603:2099207 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2098603:2099207 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2098601:2099172 [5] NCCL INFO comm 0x1b576cd0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe8a4a9e2dd502bb8 - Init COMPLETE +ip-26-0-170-143:2098601:2099208 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-170-143:2098601:2099208 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 466, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 466, in forward + out = F.linear(tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 768.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 539.94 MiB is free. Including non-PyTorch memory, this process has 78.79 GiB memory in use. Of the allocated memory 74.42 GiB is allocated by PyTorch, and 2.35 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + out = F.linear(tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 768.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 619.94 MiB is free. Including non-PyTorch memory, this process has 78.71 GiB memory in use. Of the allocated memory 74.42 GiB is allocated by PyTorch, and 2.35 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + sharded_logits = self.model( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] +hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 655, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + key_value_states = torch.cat([key_states.unsqueeze(0), value_states.unsqueeze(0)], dim=0) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 323.94 MiB is free. Including non-PyTorch memory, this process has 79.00 GiB memory in use. Of the allocated memory 75.73 GiB is allocated by PyTorch, and 1.29 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 655, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + key_value_states = torch.cat([key_states.unsqueeze(0), value_states.unsqueeze(0)], dim=0) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 323.94 MiB is free. Including non-PyTorch memory, this process has 79.00 GiB memory in use. Of the allocated memory 75.73 GiB is allocated by PyTorch, and 1.29 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + .contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 435.94 MiB is free. Including non-PyTorch memory, this process has 78.89 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.67 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 515.94 MiB is free. Including non-PyTorch memory, this process has 78.81 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.67 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( +output = model(**micro_batch) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return f(*args, **kwargs) +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return f(*args, **kwargs) +output = model(**micro_batch) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter +return forward_call(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model(sharded_logits = self.model( + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + +return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + sharded_logits = self.model( File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward +return forward_call(*args, **kwargs)return forward_call(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0]hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + return self._call_impl(*args, **kwargs) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) +return forward_call(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return forward_call(*args, **kwargs)output = self.pp_block(**new_kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + .contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 515.94 MiB is free. Including non-PyTorch memory, this process has 78.81 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.67 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward +output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward +Traceback (most recent call last): + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 655, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 539.94 MiB is free. Including non-PyTorch memory, this process has 78.79 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.60 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + key_value_states = torch.cat([key_states.unsqueeze(0), value_states.unsqueeze(0)], dim=0) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 75.73 GiB is allocated by PyTorch, and 1.48 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + .contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 515.94 MiB is free. Including non-PyTorch memory, this process has 78.81 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.67 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + .contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 435.94 MiB is free. Including non-PyTorch memory, this process has 78.89 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.67 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 655, in forward + return forward_call(*args, **kwargs)output = model(**micro_batch) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + key_value_states = torch.cat([key_states.unsqueeze(0), value_states.unsqueeze(0)], dim=0) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 75.73 GiB is allocated by PyTorch, and 1.35 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 657, in forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + key_value_states = key_value_states.permute(1, 2, 0, 3, 4).contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 76.11 GiB is allocated by PyTorch, and 1.23 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 657, in forward + key_value_states = key_value_states.permute(1, 2, 0, 3, 4).contiguous() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 203.94 MiB is free. Including non-PyTorch memory, this process has 79.12 GiB memory in use. Of the allocated memory 76.11 GiB is allocated by PyTorch, and 1.23 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[E ProcessGroupGloo.cpp:138] Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +[E ProcessGroupGloo.cpp:138] Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3375, in reduce_scatter_tensor + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3375, in reduce_scatter_tensor + work = group._reduce_scatter_base(output, input, opts) +RuntimeError: Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.170.143]:10664 + +During handling of the above exception, another exception occurred: + +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + work = group._reduce_scatter_base(output, input, opts) +RuntimeError: Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.170.143]:10664 + +During handling of the above exception, another exception occurred: + +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 469, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 469, in forward + out = differentiable_reduce_scatter_sum(out, group=group) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 146, in differentiable_reduce_scatter_sum + out = differentiable_reduce_scatter_sum(out, group=group) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 146, in differentiable_reduce_scatter_sum + return DifferentiableReduceScatterSum.apply(tensor, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return DifferentiableReduceScatterSum.apply(tensor, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 119, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 119, in forward + dist.reduce_scatter_tensor(sharded_tensor, tensor, group=group, op=dist.ReduceOp.SUM) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 43, in reduce_scatter_tensor + dist.reduce_scatter_tensor(sharded_tensor, tensor, group=group, op=dist.ReduceOp.SUM) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 43, in reduce_scatter_tensor + return dist.reduce_scatter_tensor(output=output, input=input, group=group, op=op, async_op=async_op) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 52, in wrapper + return dist.reduce_scatter_tensor(output=output, input=input, group=group, op=op, async_op=async_op) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 52, in wrapper + "args": f"{args}, {kwargs}", + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 431, in __repr__ + "args": f"{args}, {kwargs}", + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 431, in __repr__ + return torch._tensor_str._str(self, tensor_contents=tensor_contents) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 664, in _str + return torch._tensor_str._str(self, tensor_contents=tensor_contents) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 664, in _str + return _str_intern(self, tensor_contents=tensor_contents) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 595, in _str_intern + return _str_intern(self, tensor_contents=tensor_contents) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 595, in _str_intern + tensor_str = _tensor_str(self, indent) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 329, in _tensor_str + tensor_str = _tensor_str(self, indent) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 329, in _tensor_str + self = self.float() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.50 GiB. GPU 6 has a total capacty of 79.33 GiB of which 491.94 MiB is free. Including non-PyTorch memory, this process has 78.84 GiB memory in use. Of the allocated memory 75.36 GiB is allocated by PyTorch, and 1.54 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self = self.float() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.50 GiB. GPU 5 has a total capacty of 79.33 GiB of which 619.94 MiB is free. Including non-PyTorch memory, this process has 78.71 GiB memory in use. Of the allocated memory 75.36 GiB is allocated by PyTorch, and 1.42 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_3_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_2_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_3_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_2_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_3_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_2_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_3_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #30 with key NCCL_2_trace_start +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 909, in forward_with_hidden_states + fp32_sharded_logits = self.cast_to_fp32(x=sharded_logits)["output"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 876, in + module_builder=lambda: lambda x: x.float(), +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 15.97 GiB is free. Including non-PyTorch memory, this process has 63.35 GiB memory in use. Of the allocated memory 58.59 GiB is allocated by PyTorch, and 2.69 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs)return f(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 8.11 GiB is free. Including non-PyTorch memory, this process has 71.21 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.63 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 8.17 GiB is free. Including non-PyTorch memory, this process has 71.15 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.57 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 8.11 GiB is free. Including non-PyTorch memory, this process has 71.21 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.63 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 8.09 GiB is free. Including non-PyTorch memory, this process has 71.22 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.57 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 8.46 GiB is free. Including non-PyTorch memory, this process has 70.86 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.32 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward +outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + loss = self.loss( +loss = sharded_cross_entropy( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 8.77 GiB is free. Including non-PyTorch memory, this process has 70.54 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.01 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 8.65 GiB is free. Including non-PyTorch memory, this process has 70.67 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.13 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 8.32 GiB is free. Including non-PyTorch memory, this process has 71.00 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.38 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[E ProcessGroupGloo.cpp:138] Rank 3 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +[E ProcessGroupGloo.cpp:138] Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +[E ProcessGroupGloo.cpp:138] Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + trainer.train(dataloader) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model)output = model(**micro_batch) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs)return self._call_impl(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward +output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs)output = self.pp_block(**new_kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) +return self._call_impl(*args, **kwargs) File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 41, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 41, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 41, in forward + dist.all_reduce(logits_max, op=dist.ReduceOp.MAX, group=group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.all_reduce(logits_max, op=dist.ReduceOp.MAX, group=group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + dist.all_reduce(logits_max, op=dist.ReduceOp.MAX, group=group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + work = group.allreduce([tensor], opts) + work = group.allreduce([tensor], opts) +RuntimeError: Rank 3 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.170.160]:41883RuntimeError +: Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.170.160]:8388 + work = group.allreduce([tensor], opts) +RuntimeError: Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.170.160]:40068 +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 8.68 GiB is free. Including non-PyTorch memory, this process has 70.64 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.26 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward +return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1018, in forward + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + loss = self.loss( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl +return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 971, in forward + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + loss = sharded_cross_entropy( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 115, in sharded_cross_entropy + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 8.46 GiB is free. Including non-PyTorch memory, this process has 70.86 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.32 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return _ShardedCrossEntropy.apply(sharded_logits, target, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 43, in forward + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 8.51 GiB is free. Including non-PyTorch memory, this process has 70.81 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.20 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + sharded_logits = sharded_logits - logits_max.unsqueeze(dim=-1) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 16.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 8.52 GiB is free. Including non-PyTorch memory, this process has 70.79 GiB memory in use. Of the allocated memory 66.59 GiB is allocated by PyTorch, and 2.26 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[2024-12-27 17:58:53,381] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 3196831) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 17:58:53,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066372 closing signal SIGTERM +[2024-12-27 17:58:53,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066373 closing signal SIGTERM +[2024-12-27 17:58:53,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066374 closing signal SIGTERM +[2024-12-27 17:58:53,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066376 closing signal SIGTERM +[2024-12-27 17:58:53,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066377 closing signal SIGTERM +[2024-12-27 17:58:53,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066378 closing signal SIGTERM +[2024-12-27 17:58:53,388] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066379 closing signal SIGTERM +[2024-12-27 17:58:53,389] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2134478 closing signal SIGTERM +[2024-12-27 17:58:53,389] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2134479 closing signal SIGTERM +[2024-12-27 17:58:53,389] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2134481 closing signal SIGTERM +[2024-12-27 17:58:53,389] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2134482 closing signal SIGTERM +[2024-12-27 17:58:53,389] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2134483 closing signal SIGTERM +[2024-12-27 17:58:53,391] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2098596) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 17:58:53,435] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_9sytu2kk/13803565_bxtkugy6/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-132.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 3196832) + error_file: /tmp/torchelastic_9sytu2kk/13803565_bxtkugy6/attempt_0/1/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 657, in forward + key_value_states = key_value_states.permute(1, 2, 0, 3, 4).contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 76.11 GiB is allocated by PyTorch, and 1.23 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[2]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-132.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 3196833) + error_file: /tmp/torchelastic_9sytu2kk/13803565_bxtkugy6/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 655, in forward + key_value_states = torch.cat([key_states.unsqueeze(0), value_states.unsqueeze(0)], dim=0) + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 75.73 GiB is allocated by PyTorch, and 1.35 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[3]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-132.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 3196834) + error_file: /tmp/torchelastic_9sytu2kk/13803565_bxtkugy6/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 539.94 MiB is free. Including non-PyTorch memory, this process has 78.79 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.60 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[4]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-132.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 3196835) + error_file: /tmp/torchelastic_9sytu2kk/13803565_bxtkugy6/attempt_0/4/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 435.94 MiB is free. Including non-PyTorch memory, this process has 78.89 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.67 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[5]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-132.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 3196836) + error_file: /tmp/torchelastic_9sytu2kk/13803565_bxtkugy6/attempt_0/5/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 655, in forward + key_value_states = torch.cat([key_states.unsqueeze(0), value_states.unsqueeze(0)], dim=0) + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 75.73 GiB is allocated by PyTorch, and 1.48 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[6]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-132.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 3196837) + error_file: /tmp/torchelastic_9sytu2kk/13803565_bxtkugy6/attempt_0/6/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 515.94 MiB is free. Including non-PyTorch memory, this process has 78.81 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.67 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[7]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-132.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 3196838) + error_file: /tmp/torchelastic_9sytu2kk/13803565_bxtkugy6/attempt_0/7/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 515.94 MiB is free. Including non-PyTorch memory, this process has 78.81 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.67 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-132.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 3196831) + error_file: /tmp/torchelastic_9sytu2kk/13803565_bxtkugy6/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 657, in forward + key_value_states = key_value_states.permute(1, 2, 0, 3, 4).contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 203.94 MiB is free. Including non-PyTorch memory, this process has 79.12 GiB memory in use. Of the allocated memory 76.11 GiB is allocated by PyTorch, and 1.23 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +============================================================ +[2024-12-27 17:58:53,454] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_x29gvutv/13803565_58wxk5u9/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-143.ec2.internal + rank : 9 (local_rank: 1) + exitcode : 1 (pid: 2098597) + error_file: /tmp/torchelastic_x29gvutv/13803565_58wxk5u9/attempt_0/1/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 655, in forward + key_value_states = torch.cat([key_states.unsqueeze(0), value_states.unsqueeze(0)], dim=0) + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 323.94 MiB is free. Including non-PyTorch memory, this process has 79.00 GiB memory in use. Of the allocated memory 75.73 GiB is allocated by PyTorch, and 1.29 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[2]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-143.ec2.internal + rank : 10 (local_rank: 2) + exitcode : 1 (pid: 2098598) + error_file: /tmp/torchelastic_x29gvutv/13803565_58wxk5u9/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 515.94 MiB is free. Including non-PyTorch memory, this process has 78.81 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.67 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[3]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-143.ec2.internal + rank : 11 (local_rank: 3) + exitcode : 1 (pid: 2098599) + error_file: /tmp/torchelastic_x29gvutv/13803565_58wxk5u9/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 473, in forward + .contiguous() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 576.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 435.94 MiB is free. Including non-PyTorch memory, this process has 78.89 GiB memory in use. Of the allocated memory 75.17 GiB is allocated by PyTorch, and 1.67 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[4]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-143.ec2.internal + rank : 12 (local_rank: 4) + exitcode : 1 (pid: 2098600) + error_file: /tmp/torchelastic_x29gvutv/13803565_58wxk5u9/attempt_0/4/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 466, in forward + out = F.linear(tensor, weight, bias) + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 768.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 539.94 MiB is free. Including non-PyTorch memory, this process has 78.79 GiB memory in use. Of the allocated memory 74.42 GiB is allocated by PyTorch, and 2.35 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[5]: + time : 2024-12-27_17:58:27 + host : ip-26-0-170-143.ec2.internal + rank : 13 (local_rank: 5) + exitcode : 1 (pid: 2098601) + error_file: /tmp/torchelastic_x29gvutv/13803565_58wxk5u9/attempt_0/5/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3375, in reduce_scatter_tensor + work = group._reduce_scatter_base(output, input, opts) + RuntimeError: Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: + [../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.170.143]:10664 + + During handling of the above exception, another exception occurred: + + Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 469, in forward + out = differentiable_reduce_scatter_sum(out, group=group) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 146, in differentiable_reduce_scatter_sum + return DifferentiableReduceScatterSum.apply(tensor, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 119, in forward + dist.reduce_scatter_tensor(sharded_tensor, tensor, group=group, op=dist.ReduceOp.SUM) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 43, in reduce_scatter_tensor + return dist.reduce_scatter_tensor(output=output, input=input, group=group, op=op, async_op=async_op) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 52, in wrapper + "args": f"{args}, {kwargs}", + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 431, in __repr__ + return torch._tensor_str._str(self, tensor_contents=tensor_contents) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 664, in _str + return _str_intern(self, tensor_contents=tensor_contents) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 595, in _str_intern + tensor_str = _tensor_str(self, indent) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 329, in _tensor_str + self = self.float() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.50 GiB. GPU 5 has a total capacty of 79.33 GiB of which 619.94 MiB is free. Including non-PyTorch memory, this process has 78.71 GiB memory in use. Of the allocated memory 75.36 GiB is allocated by PyTorch, and 1.42 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[6]: + time : 2024-12-27_17:58:27 + host : ip-26-0-170-143.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 2098602) + error_file: /tmp/torchelastic_x29gvutv/13803565_58wxk5u9/attempt_0/6/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3375, in reduce_scatter_tensor + work = group._reduce_scatter_base(output, input, opts) + RuntimeError: Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: + [../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.170.143]:10664 + + During handling of the above exception, another exception occurred: + + Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 469, in forward + out = differentiable_reduce_scatter_sum(out, group=group) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 146, in differentiable_reduce_scatter_sum + return DifferentiableReduceScatterSum.apply(tensor, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 119, in forward + dist.reduce_scatter_tensor(sharded_tensor, tensor, group=group, op=dist.ReduceOp.SUM) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 43, in reduce_scatter_tensor + return dist.reduce_scatter_tensor(output=output, input=input, group=group, op=op, async_op=async_op) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 52, in wrapper + "args": f"{args}, {kwargs}", + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 431, in __repr__ + return torch._tensor_str._str(self, tensor_contents=tensor_contents) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 664, in _str + return _str_intern(self, tensor_contents=tensor_contents) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 595, in _str_intern + tensor_str = _tensor_str(self, indent) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor_str.py", line 329, in _tensor_str + self = self.float() + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.50 GiB. GPU 6 has a total capacty of 79.33 GiB of which 491.94 MiB is free. Including non-PyTorch memory, this process has 78.84 GiB memory in use. Of the allocated memory 75.36 GiB is allocated by PyTorch, and 1.54 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[7]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-143.ec2.internal + rank : 15 (local_rank: 7) + exitcode : 1 (pid: 2098603) + error_file: /tmp/torchelastic_x29gvutv/13803565_58wxk5u9/attempt_0/7/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 466, in forward + out = F.linear(tensor, weight, bias) + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 768.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 619.94 MiB is free. Including non-PyTorch memory, this process has 78.71 GiB memory in use. Of the allocated memory 74.42 GiB is allocated by PyTorch, and 2.35 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_17:58:26 + host : ip-26-0-170-143.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 2098596) + error_file: /tmp/torchelastic_x29gvutv/13803565_58wxk5u9/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 725, in _core_forward + output = self.attn(hidden_states=hidden_states, sequence_mask=sequence_mask) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 655, in forward + key_value_states = torch.cat([key_states.unsqueeze(0), value_states.unsqueeze(0)], dim=0) + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 323.94 MiB is free. Including non-PyTorch memory, this process has 79.00 GiB memory in use. Of the allocated memory 75.73 GiB is allocated by PyTorch, and 1.29 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +============================================================ +srun: error: ip-26-0-170-132: task 1: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13803565.0 +slurmstepd: error: *** STEP 13803565.0 ON ip-26-0-170-31 CANCELLED AT 2024-12-27T17:58:53 *** +[2024-12-27 17:58:53,707] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-27 17:58:53,707] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066372 closing signal SIGTERM +[2024-12-27 17:58:53,707] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066373 closing signal SIGTERM +[2024-12-27 17:58:53,707] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066374 closing signal SIGTERM +[2024-12-27 17:58:53,707] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066376 closing signal SIGTERM +[2024-12-27 17:58:53,707] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066377 closing signal SIGTERM +[2024-12-27 17:58:53,707] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066378 closing signal SIGTERM +[2024-12-27 17:58:53,707] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3066379 closing signal SIGTERM +[2024-12-27 17:58:53,706] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-27 17:58:53,706] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2134478 closing signal SIGTERM +[2024-12-27 17:58:53,706] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2134479 closing signal SIGTERM +[2024-12-27 17:58:53,707] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2134481 closing signal SIGTERM +[2024-12-27 17:58:53,707] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2134482 closing signal SIGTERM +[2024-12-27 17:58:53,707] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2134483 closing signal SIGTERM +srun: error: ip-26-0-170-143: task 2: Terminated +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2134404 got signal: 15 +srun: error: ip-26-0-170-160: task 3: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 3066300 got signal: 15 +srun: error: ip-26-0-170-31: task 0: Exited with exit code 1 +srun: Force Terminated StepId=13803565.0 diff --git a/logs/13803617-bench_3.57G_dp8_tp2_pp2_acc4_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13803617-bench_3.57G_dp8_tp2_pp2_acc4_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..9eaedc86a3a61960ae9fccc5462a0802d95c6ff8 --- /dev/null +++ b/logs/13803617-bench_3.57G_dp8_tp2_pp2_acc4_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,10514 @@ ++ '[' -z 13803617 ']' ++ unset FI_PROVIDER ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-166-[125,214,244],ip-26-0-167-9' ++ export 'NODELIST=ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9' ++ NODELIST='ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-166-[125,214,244],ip-26-0-167-9' ++ export MASTER_NODE=ip-26-0-166-125 ++ MASTER_NODE=ip-26-0-166-125 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ export NCCL_P2P_LEVEL=LOC ++ NCCL_P2P_LEVEL=LOC ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-43 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 52-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 52-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 52-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 52-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-166-125' +Master node: ip-26-0-166-125 ++ echo 'All nodes: ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9' +All nodes: ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 ++ echo 'World size: 32' +World size: 32 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13803617 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-166-125:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_3.57G_dp8_tp2_pp2_acc4_mbs8_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-27 18:02:56,602] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:02:56,602] torch.distributed.run: [WARNING] +[2024-12-27 18:02:56,602] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:02:56,602] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:02:56,602] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:02:56,649] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:02:56,674] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:02:56,674] torch.distributed.run: [WARNING] +[2024-12-27 18:02:56,674] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:02:56,674] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:02:56,674] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:02:56,649] torch.distributed.run: [WARNING] +[2024-12-27 18:02:56,649] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:02:56,649] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:02:56,649] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:02:56,905] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:02:56,905] torch.distributed.run: [WARNING] +[2024-12-27 18:02:56,905] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:02:56,905] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:02:56,905] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-166-125:1348305:1348305 [0] NCCL INFO Bootstrap : Using enp74s0:26.0.172.86<0> +ip-26-0-166-125:1348305:1348305 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-125:1348305:1348305 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-125:1348305:1348305 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-244:1084479:1084479 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1084478:1084478 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-125:1348307:1348307 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-125:1348308:1348308 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-125:1348309:1348309 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-125:1348312:1348312 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-125:1348306:1348306 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-125:1348311:1348311 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1084479:1084479 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-166-244:1084478:1084478 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-166-244:1084479:1084479 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1084479:1084479 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-125:1348310:1348310 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1084478:1084478 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1084478:1084478 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1080446:1080446 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1080440:1080440 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1080439:1080439 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1080443:1080443 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1080445:1080445 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1080441:1080441 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1080444:1080444 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-9:1080442:1080442 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1141900:1141900 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1141900:1141900 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-167-9:1080442:1080442 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-166-214:1141900:1141900 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1141900:1141900 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1080442:1080442 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1080442:1080442 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-125:1348307:1348307 [2] NCCL INFO Bootstrap : Using enp74s0:26.0.172.86<0> +ip-26-0-166-125:1348308:1348308 [3] NCCL INFO Bootstrap : Using enp74s0:26.0.172.86<0> +ip-26-0-166-125:1348309:1348309 [4] NCCL INFO Bootstrap : Using enp74s0:26.0.172.86<0> +ip-26-0-166-125:1348306:1348306 [1] NCCL INFO Bootstrap : Using enp74s0:26.0.172.86<0> +ip-26-0-166-125:1348312:1348312 [7] NCCL INFO Bootstrap : Using enp74s0:26.0.172.86<0> +ip-26-0-166-125:1348311:1348311 [6] NCCL INFO Bootstrap : Using enp74s0:26.0.172.86<0> +ip-26-0-166-125:1348310:1348310 [5] NCCL INFO Bootstrap : Using enp74s0:26.0.172.86<0> +ip-26-0-166-125:1348308:1348308 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-125:1348307:1348307 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-125:1348309:1348309 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-125:1348308:1348308 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-125:1348307:1348307 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-125:1348309:1348309 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-125:1348306:1348306 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-125:1348306:1348306 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-125:1348312:1348312 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-125:1348312:1348312 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-125:1348311:1348311 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-125:1348311:1348311 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-125:1348310:1348310 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-125:1348310:1348310 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1084484:1084484 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1084484:1084484 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-166-244:1084485:1084485 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1084484:1084484 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1084484:1084484 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1084485:1084485 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-166-244:1084480:1084480 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1084485:1084485 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1084485:1084485 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1080446:1080446 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-167-9:1080440:1080440 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-167-9:1080445:1080445 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-167-9:1080439:1080439 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-167-9:1080441:1080441 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-167-9:1080443:1080443 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-167-9:1080444:1080444 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.167.9<0> +ip-26-0-167-9:1080446:1080446 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1080439:1080439 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1080444:1080444 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1080440:1080440 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1080439:1080439 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1080444:1080444 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1080446:1080446 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1080440:1080440 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1084480:1084480 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-167-9:1080445:1080445 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1080445:1080445 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1080443:1080443 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1080443:1080443 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1080441:1080441 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-9:1080441:1080441 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1084480:1084480 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1084480:1084480 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1084483:1084483 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1084482:1084482 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1084483:1084483 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-166-244:1084483:1084483 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1084483:1084483 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1084482:1084482 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-166-244:1084482:1084482 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1084482:1084482 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-244:1084481:1084481 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-244:1084481:1084481 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.166.244<0> +ip-26-0-166-244:1084481:1084481 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-244:1084481:1084481 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1141899:1141899 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1141895:1141895 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1141898:1141898 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1141902:1141902 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1141897:1141897 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1141901:1141901 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1141896:1141896 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-166-214:1141899:1141899 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1141895:1141895 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1141898:1141898 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1141902:1141902 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1141901:1141901 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1141896:1141896 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1141897:1141897 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.166.214<0> +ip-26-0-166-214:1141899:1141899 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1141899:1141899 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1141896:1141896 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1141901:1141901 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1141895:1141895 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1141902:1141902 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1141895:1141895 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1141902:1141902 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1141901:1141901 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1141896:1141896 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1141898:1141898 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1141898:1141898 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-166-214:1141897:1141897 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-166-214:1141897:1141897 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/253 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO comm 0x9bd47d0 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO comm 0xa241a50 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO comm 0x9c68a60 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO comm 0x84d0900 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO comm 0xa01e6b0 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO comm 0x983a310 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO comm 0x9b71a10 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO comm 0x98f5280 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO comm 0x91466a0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO comm 0x882f370 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO comm 0x86f3ba0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO comm 0xa417c10 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO comm 0x9aa3740 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO comm 0x88a1ae0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO comm 0x8a7b7f0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO comm 0x9d25fd0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO comm 0x8d06aa0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO comm 0xa1d75f0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO comm 0xa374710 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO comm 0xa29ecf0 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO comm 0x8e011c0 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO comm 0xa267010 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO comm 0xa3533a0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO comm 0x9e5b180 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO comm 0x9fbb1f0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO comm 0x8a9a230 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO comm 0x85627e0 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO comm 0x9227be0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO comm 0x8d77010 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO comm 0x9bb4ea0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO comm 0x9c086a0 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO comm 0x845d890 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa66212fa6fadcb5 - Init START +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Trees [0] 8/24/-1->16->0 [1] 8/24/-1->16->0 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 15/-1/-1->31->-1 [3] 15/-1/-1->31->-1 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Trees [0] 16/-1/-1->0->-1 [1] 16/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Connected all rings +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Connected all rings +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Connected all rings +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Connected all rings +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Connected all rings +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO Connected all trees +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080446:1080604 [7] NCCL INFO comm 0x84d0900 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-125:1348305:1348483 [0] NCCL INFO comm 0xa01e6b0 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO Connected all trees +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141897:1142057 [2] NCCL INFO comm 0x8a7b7f0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO Connected all trees +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO Connected all trees +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348311:1348490 [6] NCCL INFO comm 0xa417c10 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO Connected all trees +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084480:1084642 [2] NCCL INFO comm 0x85627e0 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141896:1142058 [1] NCCL INFO comm 0x88a1ae0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348307:1348484 [2] NCCL INFO comm 0x9b71a10 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-214:1141901:1142060 [6] NCCL INFO comm 0x9227be0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080441:1080605 [2] NCCL INFO comm 0xa29ecf0 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-244:1084484:1084644 [6] NCCL INFO comm 0x8d06aa0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-125:1348310:1348488 [5] NCCL INFO comm 0x91466a0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-125:1348306:1348489 [1] NCCL INFO comm 0x983a310 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-214:1141900:1142056 [5] NCCL INFO comm 0x9c086a0 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-244:1084479:1084640 [1] NCCL INFO comm 0x845d890 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-167-9:1080440:1080607 [1] NCCL INFO comm 0xa374710 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-244:1084483:1084646 [5] NCCL INFO comm 0xa3533a0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080445:1080606 [6] NCCL INFO comm 0x9c68a60 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-167-9:1080444:1080603 [5] NCCL INFO comm 0x9bd47d0 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO Connected all trees +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO Connected all trees +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141902:1142062 [7] NCCL INFO comm 0x8d77010 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1141898:1142059 [3] NCCL INFO comm 0x9d25fd0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348308:1348485 [3] NCCL INFO comm 0x86f3ba0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-125:1348312:1348486 [7] NCCL INFO comm 0x882f370 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-125:1348309:1348487 [4] NCCL INFO comm 0x98f5280 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-214:1141899:1142061 [4] NCCL INFO comm 0x9bb4ea0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-214:1141895:1142063 [0] NCCL INFO comm 0x9aa3740 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-167-9:1080442:1080601 [3] NCCL INFO comm 0xa267010 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084485:1084643 [7] NCCL INFO comm 0xa1d75f0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084481:1084647 [3] NCCL INFO comm 0x8a9a230 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084478:1084641 [0] NCCL INFO comm 0x9fbb1f0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-167-9:1080443:1080602 [4] NCCL INFO comm 0xa241a50 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-167-9:1080439:1080608 [0] NCCL INFO comm 0x8e011c0 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-244:1084482:1084645 [4] NCCL INFO comm 0x9e5b180 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfa66212fa6fadcb5 - Init COMPLETE +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO comm 0x8a66570 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO comm 0xa64ff40 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO comm 0x8cb3d20 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO comm 0x8adaaa0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO comm 0x9f5eba0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO comm 0x9cda9f0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO comm 0x937f740 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO comm 0x9b2dd00 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO comm 0x892bfe0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO comm 0x8707b30 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO comm 0x9da8830 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO comm 0x9a70830 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO comm 0xa25da80 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO comm 0x9ea1580 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO comm 0x9e0d200 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO comm 0xa479ec0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO comm 0xa49d720 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO comm 0xa411220 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO comm 0xa4d5e40 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO comm 0xa5ab6f0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO comm 0x9036d40 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO comm 0x8f3eec0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO comm 0xa58b3e0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO comm 0x8cd2750 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO comm 0x879a640 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO comm 0xa093d20 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO comm 0x8695d20 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO comm 0x8fb01e0 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO comm 0xa1f3040 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO comm 0x9460880 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO comm 0x9debfb0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO comm 0x9e41160 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4be3bbc2df2bfc11 - Init START +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Trees [0] 16/-1/-1->0->-1 [1] 16/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 15/-1/-1->31->-1 [3] 15/-1/-1->31->-1 +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Trees [0] 8/24/-1->16->0 [1] 8/24/-1->16->0 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 01/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 02/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 03/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Connected all rings +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Connected all rings +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Connected all rings +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Connected all rings +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Connected all rings +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Connected all rings +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Connected all rings +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Connected all rings +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Connected all rings +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 01/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 01/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO Connected all trees +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348305:1348721 [0] NCCL INFO comm 0xa25da80 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-167-9:1080446:1080831 [7] NCCL INFO comm 0x8707b30 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141897:1142289 [2] NCCL INFO comm 0x8cb3d20 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141901:1142292 [6] NCCL INFO comm 0x9460880 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-214:1141896:1142288 [1] NCCL INFO comm 0x8adaaa0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO Connected all trees +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO Connected all trees +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO Connected all trees +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348311:1348727 [6] NCCL INFO comm 0xa64ff40 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348306:1348722 [1] NCCL INFO comm 0x9a70830 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-125:1348307:1348723 [2] NCCL INFO comm 0x9da8830 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-244:1084484:1084872 [6] NCCL INFO comm 0x8f3eec0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080445:1080832 [6] NCCL INFO comm 0x9ea1580 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-244:1084480:1084877 [2] NCCL INFO comm 0x879a640 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-167-9:1080441:1080836 [2] NCCL INFO comm 0xa4d5e40 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142290 [3] NCCL INFO comm 0x9f5eba0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO Connected all trees +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO Connected all trees +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141900:1142293 [5] NCCL INFO comm 0x9e41160 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141899:1142291 [4] NCCL INFO comm 0x9debfb0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-214:1141902:1142294 [7] NCCL INFO comm 0x8fb01e0 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-244:1084481:1084875 [3] NCCL INFO comm 0x8cd2750 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-244:1084485:1084871 [7] NCCL INFO comm 0xa411220 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO Connected all trees +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348312:1348728 [7] NCCL INFO comm 0x8a66570 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO Connected all trees +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348308:1348724 [3] NCCL INFO comm 0x892bfe0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141895:1142287 [0] NCCL INFO comm 0x9cda9f0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348310:1348726 [5] NCCL INFO comm 0x937f740 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-244:1084479:1084876 [1] NCCL INFO comm 0x8695d20 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-244:1084478:1084878 [0] NCCL INFO comm 0xa1f3040 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-167-9:1080440:1080837 [1] NCCL INFO comm 0xa5ab6f0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-125:1348309:1348725 [4] NCCL INFO comm 0x9b2dd00 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-244:1084483:1084874 [5] NCCL INFO comm 0xa58b3e0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-166-244:1084482:1084873 [4] NCCL INFO comm 0xa093d20 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-167-9:1080442:1080835 [3] NCCL INFO comm 0xa49d720 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080444:1080833 [5] NCCL INFO comm 0x9e0d200 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-167-9:1080443:1080834 [4] NCCL INFO comm 0xa479ec0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +ip-26-0-167-9:1080439:1080838 [0] NCCL INFO comm 0x9036d40 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4be3bbc2df2bfc11 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO comm 0xa423ee0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x612260b82252f4f0 - Init START +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO comm 0xa205d90 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x612260b82252f4f0 - Init START +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO comm 0x86a89e0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x612260b82252f4f0 - Init START +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO comm 0x87ad300 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x612260b82252f4f0 - Init START +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO comm 0x8f51b80 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x612260b82252f4f0 - Init START +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO comm 0xa59e0a0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x612260b82252f4f0 - Init START +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO comm 0x8ce5410 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x612260b82252f4f0 - Init START +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO comm 0xa0a69e0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x612260b82252f4f0 - Init START +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO comm 0xa48cb80 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2d43ef346ee0d6b - Init START +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO comm 0xa4b1a40 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2d43ef346ee0d6b - Init START +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO comm 0xa5becc0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2d43ef346ee0d6b - Init START +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO comm 0x871ab80 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2d43ef346ee0d6b - Init START +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO comm 0x904c050 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2d43ef346ee0d6b - Init START +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO comm 0xa4e9080 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2d43ef346ee0d6b - Init START +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO comm 0x9e1fec0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2d43ef346ee0d6b - Init START +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO comm 0x9eb4300 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2d43ef346ee0d6b - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO comm 0x8aed760 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xecc68d8bd02e765c - Init START +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO comm 0x9cee900 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xecc68d8bd02e765c - Init START +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO comm 0x8fc2ea0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xecc68d8bd02e765c - Init START +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO comm 0x893eca0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xce001990d5de77bf - Init START +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO comm 0x9473540 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xecc68d8bd02e765c - Init START +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO comm 0x9dff600 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xecc68d8bd02e765c - Init START +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO comm 0x8cc69e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xecc68d8bd02e765c - Init START +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO comm 0x9e53e20 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xecc68d8bd02e765c - Init START +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO comm 0x9f71860 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xecc68d8bd02e765c - Init START +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO comm 0x9dbbd90 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xce001990d5de77bf - Init START +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO comm 0x9b409c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xce001990d5de77bf - Init START +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO comm 0x9a84d90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xce001990d5de77bf - Init START +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO comm 0x9392400 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xce001990d5de77bf - Init START +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO comm 0xa270740 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xce001990d5de77bf - Init START +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO comm 0x8a7a310 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xce001990d5de77bf - Init START +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO comm 0xa662c00 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xce001990d5de77bf - Init START +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Connected all rings +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Connected all rings +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Connected all rings +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Connected all rings +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Connected all rings +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Connected all rings +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Connected all rings +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Connected all rings +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Connected all rings +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO Connected all trees +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348312:1348761 [7] NCCL INFO comm 0x8a7a310 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xce001990d5de77bf - Init COMPLETE +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO Connected all trees +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO Connected all trees +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Connected all rings +ip-26-0-166-125:1348305:1348757 [0] NCCL INFO comm 0xa270740 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xce001990d5de77bf - Init COMPLETE +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO Connected all trees +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348306:1348763 [1] NCCL INFO comm 0x9a84d90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xce001990d5de77bf - Init COMPLETE +ip-26-0-167-9:1080446:1080875 [7] NCCL INFO comm 0x871ab80 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2d43ef346ee0d6b - Init COMPLETE +ip-26-0-166-125:1348307:1348760 [2] NCCL INFO comm 0x9dbbd90 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xce001990d5de77bf - Init COMPLETE +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO Connected all trees +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348311:1348762 [6] NCCL INFO comm 0xa662c00 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xce001990d5de77bf - Init COMPLETE +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO Connected all trees +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080439:1080868 [0] NCCL INFO comm 0x904c050 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2d43ef346ee0d6b - Init COMPLETE +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO Connected all trees +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO Connected all trees +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348308:1348759 [3] NCCL INFO comm 0x893eca0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xce001990d5de77bf - Init COMPLETE +ip-26-0-166-125:1348310:1348764 [5] NCCL INFO comm 0x9392400 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xce001990d5de77bf - Init COMPLETE +ip-26-0-166-125:1348309:1348758 [4] NCCL INFO comm 0x9b409c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xce001990d5de77bf - Init COMPLETE +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080874 [1] NCCL INFO comm 0xa5becc0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2d43ef346ee0d6b - Init COMPLETE +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080872 [6] NCCL INFO comm 0x9eb4300 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2d43ef346ee0d6b - Init COMPLETE +ip-26-0-167-9:1080441:1080873 [2] NCCL INFO comm 0xa4e9080 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2d43ef346ee0d6b - Init COMPLETE +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080442:1080871 [3] NCCL INFO comm 0xa4b1a40 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2d43ef346ee0d6b - Init COMPLETE +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084485:1084907 [7] NCCL INFO comm 0xa423ee0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x612260b82252f4f0 - Init COMPLETE +ip-26-0-166-244:1084478:1084906 [0] NCCL INFO comm 0xa205d90 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x612260b82252f4f0 - Init COMPLETE +ip-26-0-167-9:1080444:1080869 [5] NCCL INFO comm 0x9e1fec0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2d43ef346ee0d6b - Init COMPLETE +ip-26-0-167-9:1080443:1080870 [4] NCCL INFO comm 0xa48cb80 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2d43ef346ee0d6b - Init COMPLETE +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084484:1084911 [6] NCCL INFO comm 0x8f51b80 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x612260b82252f4f0 - Init COMPLETE +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084479:1084909 [1] NCCL INFO comm 0x86a89e0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x612260b82252f4f0 - Init COMPLETE +ip-26-0-166-244:1084483:1084910 [5] NCCL INFO comm 0xa59e0a0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x612260b82252f4f0 - Init COMPLETE +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084481:1084908 [3] NCCL INFO comm 0x8ce5410 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x612260b82252f4f0 - Init COMPLETE +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141895:1142320 [0] NCCL INFO comm 0x9cee900 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xecc68d8bd02e765c - Init COMPLETE +ip-26-0-166-214:1141902:1142325 [7] NCCL INFO comm 0x8fc2ea0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xecc68d8bd02e765c - Init COMPLETE +ip-26-0-166-244:1084480:1084912 [2] NCCL INFO comm 0x87ad300 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x612260b82252f4f0 - Init COMPLETE +ip-26-0-166-244:1084482:1084913 [4] NCCL INFO comm 0xa0a69e0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x612260b82252f4f0 - Init COMPLETE +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141896:1142326 [1] NCCL INFO comm 0x8aed760 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xecc68d8bd02e765c - Init COMPLETE +ip-26-0-166-214:1141901:1142321 [6] NCCL INFO comm 0x9473540 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xecc68d8bd02e765c - Init COMPLETE +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142322 [3] NCCL INFO comm 0x9f71860 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xecc68d8bd02e765c - Init COMPLETE +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141897:1142327 [2] NCCL INFO comm 0x8cc69e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xecc68d8bd02e765c - Init COMPLETE +ip-26-0-166-214:1141900:1142323 [5] NCCL INFO comm 0x9e53e20 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xecc68d8bd02e765c - Init COMPLETE +ip-26-0-166-214:1141899:1142324 [4] NCCL INFO comm 0x9dff600 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xecc68d8bd02e765c - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO comm 0xc1aa5f0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9b972f79c744eb69 - Init START +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO comm 0xbb3d320 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9b972f79c744eb69 - Init START +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO comm 0xaf2a040 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9b058ca47e1153f3 - Init START +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO comm 0xb710ee0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9b058ca47e1153f3 - Init START +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO comm 0xc1cfc70 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd20d2d342845b249 - Init START +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO comm 0xb7c0c00 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd20d2d342845b249 - Init START +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO comm 0x9f8fd40 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf08c0070932ba98b - Init START +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO comm 0xb192bc0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf08c0070932ba98b - Init START +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO comm 0xb895ef0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1e0b5e1cbbbd14e0 - Init START +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO comm 0xa324d40 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1e0b5e1cbbbd14e0 - Init START +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO comm 0xafe2a00 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb72fc6a8d9ad9a3f - Init START +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO comm 0xa834f60 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb72fc6a8d9ad9a3f - Init START +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO comm 0xbb05560 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7aa6a518b24be8df - Init START +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO comm 0x9f1d760 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7aa6a518b24be8df - Init START +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO comm 0xbbd12e0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb5d03bceaa2dae17 - Init START +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO comm 0xa437dd0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb5d03bceaa2dae17 - Init START +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO comm 0xbdc4020 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd85bfb0242640e0e - Init START +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO comm 0xc2bb2f0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd85bfb0242640e0e - Init START +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO comm 0xa465ae0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x89b242f53f69d5b - Init START +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO comm 0xa9165a0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x89b242f53f69d5b - Init START +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO comm 0xb2f5ce0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x86252e85a896b940 - Init START +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO comm 0xb2a4a00 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x86252e85a896b940 - Init START +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO comm 0xb415450 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9c59258fb7e4585 - Init START +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO comm 0xa16b520 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9c59258fb7e4585 - Init START +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO comm 0xc141180 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1ad024c16e9f147c - Init START +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO comm 0xa229f40 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1ad024c16e9f147c - Init START +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO comm 0x9982390 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x86586d53e482ae51 - Init START +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO comm 0xb4e0760 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x86586d53e482ae51 - Init START +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO comm 0x9fbdd40 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf75a54c182258b00 - Init START +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO comm 0xa4cbb40 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf75a54c182258b00 - Init START +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO comm 0x9de18d0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7b62bc5b1248acfb - Init START +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO comm 0xb25f320 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7b62bc5b1248acfb - Init START +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Connected all rings +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO Connected all trees +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Connected all rings +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO Connected all trees +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348308:1348807 [3] NCCL INFO comm 0x9de18d0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7b62bc5b1248acfb - Init COMPLETE +ip-26-0-166-125:1348307:1348806 [2] NCCL INFO comm 0xb25f320 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7b62bc5b1248acfb - Init COMPLETE +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Connected all rings +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO Connected all trees +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141899:1142366 [4] NCCL INFO comm 0xb2a4a00 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x86252e85a896b940 - Init COMPLETE +ip-26-0-166-214:1141900:1142369 [5] NCCL INFO comm 0xb2f5ce0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x86252e85a896b940 - Init COMPLETE +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Connected all rings +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO Connected all trees +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Connected all rings +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO Connected all trees +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Connected all rings +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO Connected all trees +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348306:1348801 [1] NCCL INFO comm 0xaf2a040 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9b058ca47e1153f3 - Init COMPLETE +ip-26-0-166-125:1348305:1348798 [0] NCCL INFO comm 0xb710ee0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9b058ca47e1153f3 - Init COMPLETE +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Connected all rings +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Connected all rings +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO Connected all trees +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO Connected all trees +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348310:1348803 [5] NCCL INFO comm 0xa834f60 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb72fc6a8d9ad9a3f - Init COMPLETE +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141897:1142365 [2] NCCL INFO comm 0xa16b520 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x9c59258fb7e4585 - Init COMPLETE +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348309:1348800 [4] NCCL INFO comm 0xafe2a00 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb72fc6a8d9ad9a3f - Init COMPLETE +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142360 [0] NCCL INFO comm 0xb192bc0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf08c0070932ba98b - Init COMPLETE +ip-26-0-166-214:1141896:1142363 [1] NCCL INFO comm 0x9f8fd40 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf08c0070932ba98b - Init COMPLETE +ip-26-0-166-214:1141898:1142367 [3] NCCL INFO comm 0xb415450 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x9c59258fb7e4585 - Init COMPLETE +ip-26-0-166-214:1141901:1142364 [6] NCCL INFO comm 0xa9165a0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x89b242f53f69d5b - Init COMPLETE +ip-26-0-166-214:1141902:1142368 [7] NCCL INFO comm 0xa465ae0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x89b242f53f69d5b - Init COMPLETE +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348312:1348804 [7] NCCL INFO comm 0x9f1d760 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7aa6a518b24be8df - Init COMPLETE +ip-26-0-166-125:1348311:1348802 [6] NCCL INFO comm 0xbb05560 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7aa6a518b24be8df - Init COMPLETE +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 00/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 01/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 00/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 01/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080443:1080909 [4] NCCL INFO comm 0xc1aa5f0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9b972f79c744eb69 - Init COMPLETE +ip-26-0-167-9:1080444:1080914 [5] NCCL INFO comm 0xbb3d320 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9b972f79c744eb69 - Init COMPLETE +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO comm 0xc1c4ce0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xdad43ecad8e5929c - Init START +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO comm 0xbce9630 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xdad43ecad8e5929c - Init START +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO comm 0xbb57d90 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbbc5a018bbecfc25 - Init START +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO comm 0xbd3ab40 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbbc5a018bbecfc25 - Init START +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084485:1084953 [7] NCCL INFO comm 0xc141180 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1ad024c16e9f147c - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084484:1084952 [6] NCCL INFO comm 0xa229f40 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1ad024c16e9f147c - Init COMPLETE +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO comm 0xc15bf30 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd130adc8a2e3a8 - Init START +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO comm 0xa962750 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd130adc8a2e3a8 - Init START +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO comm 0xac6f170 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8621a5fb5cf09023 - Init START +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO comm 0xc54a310 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8621a5fb5cf09023 - Init START +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084481:1084951 [3] NCCL INFO comm 0x9fbdd40 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf75a54c182258b00 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO comm 0xaa02f30 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb9e63ddb78a157 - Init START +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO comm 0xa826830 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb9e63ddb78a157 - Init START +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084480:1084950 [2] NCCL INFO comm 0xa4cbb40 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf75a54c182258b00 - Init COMPLETE +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO comm 0xbca44d0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x36a5138e7f43e26d - Init START +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO comm 0xa4e5280 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x36a5138e7f43e26d - Init START +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1084947 [1] NCCL INFO comm 0x9982390 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x86586d53e482ae51 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO comm 0xa3c7530 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5aa75a29174ccb6 - Init START +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO comm 0xb96e9a0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5aa75a29174ccb6 - Init START +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084483:1084949 [5] NCCL INFO comm 0xc2bb2f0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd85bfb0242640e0e - Init COMPLETE +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO comm 0xc2d5d10 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xba071bab444424c7 - Init START +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO comm 0xb279a60 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xba071bab444424c7 - Init START +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084478:1084944 [0] NCCL INFO comm 0xb4e0760 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x86586d53e482ae51 - Init COMPLETE +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO comm 0xbf25940 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x63efd9b9226ae843 - Init START +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO comm 0xc155d10 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x63efd9b9226ae843 - Init START +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084482:1084946 [4] NCCL INFO comm 0xbdc4020 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd85bfb0242640e0e - Init COMPLETE +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO comm 0xbddef30 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf3eca97c86f25391 - Init START +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO comm 0xba27950 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf3eca97c86f25391 - Init START +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Connected all rings +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080446:1080912 [7] NCCL INFO comm 0xa437dd0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb5d03bceaa2dae17 - Init COMPLETE +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO comm 0xa452fb0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x629478205aa79978 - Init START +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO comm 0xaeaaa30 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x629478205aa79978 - Init START +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080445:1080908 [6] NCCL INFO comm 0xbbd12e0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb5d03bceaa2dae17 - Init COMPLETE +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO comm 0xbbebef0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd0c5cc61e86adba - Init START +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO comm 0xb35b390 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd0c5cc61e86adba - Init START +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080439:1080910 [0] NCCL INFO comm 0xa324d40 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1e0b5e1cbbbd14e0 - Init COMPLETE +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO comm 0xad69cd0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc94929117a9ef16b - Init START +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO comm 0xbbd7c00 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc94929117a9ef16b - Init START +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080440:1080915 [1] NCCL INFO comm 0xb895ef0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1e0b5e1cbbbd14e0 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO comm 0xc2db060 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x260972fefb5b41ac - Init START +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO comm 0xa9d4eb0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x260972fefb5b41ac - Init START +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-9:1080442:1080913 [3] NCCL INFO comm 0xc1cfc70 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xd20d2d342845b249 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO comm 0xc1eaf10 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x69d09617c46425b4 - Init START +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO comm 0xbe5a310 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x69d09617c46425b4 - Init START +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080441:1080911 [2] NCCL INFO comm 0xb7c0c00 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd20d2d342845b249 - Init COMPLETE +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO comm 0xc205a40 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x92f0872425f64720 - Init START +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO comm 0xabb0380 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x92f0872425f64720 - Init START +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080443:1080934 [4] NCCL INFO comm 0xc1c4ce0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xdad43ecad8e5929c - Init COMPLETE +ip-26-0-167-9:1080444:1080935 [5] NCCL INFO comm 0xbb57d90 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbbc5a018bbecfc25 - Init COMPLETE +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Connected all rings +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO Connected all trees +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348311:1348831 [6] NCCL INFO comm 0xc54a310 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8621a5fb5cf09023 - Init COMPLETE +ip-26-0-166-244:1084484:1084974 [6] NCCL INFO comm 0xac6f170 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8621a5fb5cf09023 - Init COMPLETE +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Connected all rings +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO Connected all trees +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084485:1084973 [7] NCCL INFO comm 0xc15bf30 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd130adc8a2e3a8 - Init COMPLETE +ip-26-0-166-125:1348312:1348829 [7] NCCL INFO comm 0xa962750 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd130adc8a2e3a8 - Init COMPLETE +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084481:1084975 [3] NCCL INFO comm 0xaa02f30 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb9e63ddb78a157 - Init COMPLETE +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Connected all rings +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO Connected all trees +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348833 [3] NCCL INFO comm 0xa826830 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb9e63ddb78a157 - Init COMPLETE +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Connected all rings +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080941 [6] NCCL INFO comm 0xbbebef0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd0c5cc61e86adba - Init COMPLETE +ip-26-0-167-9:1080446:1080940 [7] NCCL INFO comm 0xa452fb0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x629478205aa79978 - Init COMPLETE +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Connected all rings +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO Connected all trees +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084480:1084976 [2] NCCL INFO comm 0xa4e5280 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x36a5138e7f43e26d - Init COMPLETE +ip-26-0-167-9:1080439:1080942 [0] NCCL INFO comm 0xad69cd0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc94929117a9ef16b - Init COMPLETE +ip-26-0-166-125:1348307:1348835 [2] NCCL INFO comm 0xbca44d0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x36a5138e7f43e26d - Init COMPLETE +ip-26-0-167-9:1080440:1080943 [1] NCCL INFO comm 0xc2db060 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x260972fefb5b41ac - Init COMPLETE +ip-26-0-166-214:1141899:1142389 [4] NCCL INFO comm 0xbce9630 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xdad43ecad8e5929c - Init COMPLETE +ip-26-0-166-214:1141900:1142391 [5] NCCL INFO comm 0xbd3ab40 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbbc5a018bbecfc25 - Init COMPLETE +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Connected all rings +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO Connected all trees +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084479:1084977 [1] NCCL INFO comm 0xa3c7530 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5aa75a29174ccb6 - Init COMPLETE +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Connected all rings +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO Connected all trees +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Connected all rings +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO Connected all trees +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348306:1348837 [1] NCCL INFO comm 0xb96e9a0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5aa75a29174ccb6 - Init COMPLETE +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Connected all rings +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO Connected all trees +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141902:1142398 [7] NCCL INFO comm 0xaeaaa30 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x629478205aa79978 - Init COMPLETE +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348309:1348844 [4] NCCL INFO comm 0xba27950 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf3eca97c86f25391 - Init COMPLETE +ip-26-0-166-125:1348305:1348842 [0] NCCL INFO comm 0xc155d10 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x63efd9b9226ae843 - Init COMPLETE +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1084478:1084979 [0] NCCL INFO comm 0xbf25940 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x63efd9b9226ae843 - Init COMPLETE +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348310:1348840 [5] NCCL INFO comm 0xb279a60 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xba071bab444424c7 - Init COMPLETE +ip-26-0-166-244:1084483:1084978 [5] NCCL INFO comm 0xc2d5d10 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xba071bab444424c7 - Init COMPLETE +ip-26-0-166-244:1084482:1084980 [4] NCCL INFO comm 0xbddef30 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf3eca97c86f25391 - Init COMPLETE +ip-26-0-166-214:1141896:1142404 [1] NCCL INFO comm 0xa9d4eb0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x260972fefb5b41ac - Init COMPLETE +ip-26-0-166-214:1141901:1142400 [6] NCCL INFO comm 0xb35b390 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd0c5cc61e86adba - Init COMPLETE +ip-26-0-166-214:1141895:1142402 [0] NCCL INFO comm 0xbbd7c00 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc94929117a9ef16b - Init COMPLETE +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080442:1080945 [3] NCCL INFO comm 0xc1eaf10 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x69d09617c46425b4 - Init COMPLETE +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142406 [3] NCCL INFO comm 0xbe5a310 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x69d09617c46425b4 - Init COMPLETE +ip-26-0-166-214:1141897:1142414 [2] NCCL INFO comm 0xabb0380 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x92f0872425f64720 - Init COMPLETE +ip-26-0-167-9:1080441:1080953 [2] NCCL INFO comm 0xc205a40 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x92f0872425f64720 - Init COMPLETE +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO comm 0xbdbf400 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8a730d2677615cb8 - Init START +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO comm 0xc3975e0 rank 6 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8a730d2677615cb8 - Init START +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO comm 0xc3f2e50 rank 5 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8a730d2677615cb8 - Init START +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO comm 0xaf579a0 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8a730d2677615cb8 - Init START +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO comm 0xae5bea0 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8a730d2677615cb8 - Init START +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO comm 0xc1130c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8a730d2677615cb8 - Init START +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO comm 0xbfb1330 rank 2 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8a730d2677615cb8 - Init START +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO comm 0xa6b94f0 rank 1 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8a730d2677615cb8 - Init START +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO comm 0xc575280 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x303d6fecc3d2249c - Init START +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO comm 0xba52570 rank 2 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x303d6fecc3d2249c - Init START +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO comm 0xbccf9b0 rank 1 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x303d6fecc3d2249c - Init START +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO comm 0xc186650 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x303d6fecc3d2249c - Init START +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO comm 0xbd14d00 rank 6 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x303d6fecc3d2249c - Init START +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO comm 0xb386880 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x303d6fecc3d2249c - Init START +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO comm 0xabdb4c0 rank 5 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x303d6fecc3d2249c - Init START +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO comm 0xbc032d0 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x303d6fecc3d2249c - Init START +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO comm 0xa98d500 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98423fabe9e87501 - Init START +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO comm 0xb99b0a0 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x98423fabe9e87501 - Init START +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO comm 0xb2a42c0 rank 2 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x98423fabe9e87501 - Init START +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO comm 0xa850e10 rank 1 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98423fabe9e87501 - Init START +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO comm 0xaed5040 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98423fabe9e87501 - Init START +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO comm 0xbd658c0 rank 6 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x98423fabe9e87501 - Init START +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO comm 0xbe84f60 rank 5 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98423fabe9e87501 - Init START +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO comm 0xa9fff80 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x98423fabe9e87501 - Init START +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO comm 0xa5b44c0 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfdb804b5f0cc73a2 - Init START +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO comm 0xc32e200 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfdb804b5f0cc73a2 - Init START +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO comm 0xabefe80 rank 1 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfdb804b5f0cc73a2 - Init START +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO comm 0xc4a7eb0 rank 2 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfdb804b5f0cc73a2 - Init START +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO comm 0xa625110 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfdb804b5f0cc73a2 - Init START +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO comm 0xbd2a130 rank 6 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfdb804b5f0cc73a2 - Init START +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO comm 0xc3bcba0 rank 5 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfdb804b5f0cc73a2 - Init START +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO comm 0xc4c7d50 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfdb804b5f0cc73a2 - Init START +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 00/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 01/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 02/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 03/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 00/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 01/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 02/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 03/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 00/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 00/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 01/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 02/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 01/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 02/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 03/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 03/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 00/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 01/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 02/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 03/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 00/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 01/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 02/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 03/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 00/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 01/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 02/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 03/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 00/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 01/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 02/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 03/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 00/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 00/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 01/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 02/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 01/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 03/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 02/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 00/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 03/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 01/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 00/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 02/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 01/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 03/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 02/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 03/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 00/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 00/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 00/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 01/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 00/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 02/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 01/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 01/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 02/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 01/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 03/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 03/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 02/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 00/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 02/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 01/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 03/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 03/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 02/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 00/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 00/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 03/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 01/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 02/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 03/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 01/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 02/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 03/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 00/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 00/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 01/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 02/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 01/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 03/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 02/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 00/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 03/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 00/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 01/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 01/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 02/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 02/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 03/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 03/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 00/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 01/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 02/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 03/0 : 1[2] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 00/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 00/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 01/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 02/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 01/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 03/0 : 5[2] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 02/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 03/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 00/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 00/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 01/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 00/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 01/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 02/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 02/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 03/0 : 6[4] -> 7[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 01/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 00/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 03/0 : 4[0] -> 5[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 01/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 02/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 03/0 : 6[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 02/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 03/0 : 4[0] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 00/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 01/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 00/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 02/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 01/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 03/0 : 7[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 02/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 03/0 : 5[2] -> 6[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 00/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 01/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 02/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 03/0 : 2[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 00/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 00/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 01/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 01/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 02/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 02/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 03/0 : 3[6] -> 4[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 03/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 00/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 01/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 02/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 03/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 00/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 01/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 02/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 03/0 : 4[1] -> 5[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 00/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 01/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 02/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 03/0 : 4[1] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 00/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 00/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 01/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 01/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 02/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 02/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 00/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 03/0 : 5[3] -> 6[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 03/0 : 1[3] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 01/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 02/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 00/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 00/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 03/0 : 6[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 01/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 01/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 02/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 03/0 : 2[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 02/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 00/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 03/0 : 5[3] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 01/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 00/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 01/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 02/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 02/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 03/0 : 3[7] -> 4[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 03/0 : 6[5] -> 7[7] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 00/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 01/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 02/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 03/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 00/0 : 2[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 01/0 : 2[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 02/0 : 3[6] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 03/0 : 3[6] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 00/0 : 4[0] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 01/0 : 4[0] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 00/0 : 2[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 01/0 : 2[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 00/0 : 4[0] -> 6[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 01/0 : 4[0] -> 6[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 02/0 : 1[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 03/0 : 1[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 02/0 : 3[6] -> 5[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 00/0 : 4[0] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 00/0 : 6[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 03/0 : 3[6] -> 5[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 01/0 : 4[0] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 01/0 : 6[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 02/0 : 3[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 03/0 : 3[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 02/0 : 5[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 03/0 : 5[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 02/0 : 5[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 00/0 : 7[6] -> 6[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 03/0 : 5[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Channel 01/0 : 7[6] -> 6[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 00/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 01/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 00/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 02/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 01/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 03/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 02/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 02/0 : 5[2] -> 4[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 03/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Channel 03/0 : 5[2] -> 4[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 00/0 : 6[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 01/0 : 6[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 00/0 : 4[0] -> 2[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 01/0 : 4[0] -> 2[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 00/0 : 7[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 01/0 : 7[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 00/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 02/0 : 5[2] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 00/0 : 3[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 01/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Channel 03/0 : 5[2] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 01/0 : 3[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 02/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Channel 03/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 00/0 : 2[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 01/0 : 2[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 02/0 : 1[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 03/0 : 1[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 02/0 : 3[7] -> 5[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 03/0 : 3[7] -> 5[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 02/0 : 3[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 03/0 : 3[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Connected all rings +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 00/0 : 2[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 01/0 : 2[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 00/0 : 4[1] -> 6[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 01/0 : 4[1] -> 6[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 00/0 : 4[1] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 01/0 : 4[1] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 02/0 : 3[7] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 03/0 : 3[7] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 02/0 : 5[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 03/0 : 5[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 02/0 : 5[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 03/0 : 5[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 00/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 01/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 02/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 00/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 03/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 01/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 02/0 : 5[3] -> 4[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 02/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Channel 03/0 : 5[3] -> 4[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 03/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 00/0 : 4[1] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 01/0 : 4[1] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 00/0 : 6[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 01/0 : 6[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 00/0 : 6[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 01/0 : 6[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 00/0 : 4[1] -> 2[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 01/0 : 4[1] -> 2[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 00/0 : 7[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 01/0 : 7[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 00/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 02/0 : 5[3] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 01/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Channel 03/0 : 5[3] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 00/0 : 3[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 02/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 01/0 : 3[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Channel 03/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Connected all rings +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Connected all rings +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Connected all rings +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 00/0 : 2[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 01/0 : 2[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1080961 [6] NCCL INFO comm 0xbdbf400 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8a730d2677615cb8 - Init COMPLETE +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084479:1085004 [1] NCCL INFO comm 0xa5b44c0 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfdb804b5f0cc73a2 - Init COMPLETE +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Connected all rings +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 00/0 : 4[0] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 02/0 : 1[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 01/0 : 4[0] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 03/0 : 1[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 02/0 : 3[6] -> 5[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 03/0 : 3[6] -> 5[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 02/0 : 3[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 03/0 : 3[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 02/0 : 3[6] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 03/0 : 3[6] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080446:1080966 [7] NCCL INFO comm 0xa625110 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfdb804b5f0cc73a2 - Init COMPLETE +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 02/0 : 5[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 03/0 : 5[2] -> 3[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084478:1084999 [0] NCCL INFO comm 0xc1130c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8a730d2677615cb8 - Init COMPLETE +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084480:1085001 [2] NCCL INFO comm 0xa6b94f0 rank 1 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8a730d2677615cb8 - Init COMPLETE +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Connected all rings +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Connected all rings +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Connected all rings +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 00/0 : 2[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 01/0 : 2[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085005 [5] NCCL INFO comm 0xc4a7eb0 rank 2 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfdb804b5f0cc73a2 - Init COMPLETE +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Connected all rings +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 02/0 : 1[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 03/0 : 1[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 02/0 : 3[7] -> 5[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 03/0 : 3[7] -> 5[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 02/0 : 3[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 03/0 : 3[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 00/0 : 2[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 01/0 : 2[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 00/0 : 4[0] -> 6[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 01/0 : 4[0] -> 6[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 00/0 : 4[0] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 01/0 : 4[0] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 00/0 : 6[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 01/0 : 6[4] -> 4[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 00/0 : 6[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 01/0 : 6[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 00/0 : 4[0] -> 2[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 01/0 : 4[0] -> 2[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 00/0 : 7[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 01/0 : 7[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 00/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 01/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 02/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Channel 03/0 : 6[4] -> 5[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 00/0 : 7[6] -> 6[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 02/0 : 5[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Channel 01/0 : 7[6] -> 6[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 03/0 : 5[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 00/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 00/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 01/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 01/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 02/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 02/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 03/0 : 6[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 03/0 : 2[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 02/0 : 5[2] -> 4[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Channel 03/0 : 5[2] -> 4[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 00/0 : 3[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 02/0 : 5[2] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Channel 03/0 : 5[2] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 01/0 : 3[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084481:1085007 [3] NCCL INFO comm 0xabefe80 rank 1 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfdb804b5f0cc73a2 - Init COMPLETE +ip-26-0-167-9:1080444:1080964 [5] NCCL INFO comm 0xbd2a130 rank 6 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfdb804b5f0cc73a2 - Init COMPLETE +ip-26-0-167-9:1080442:1080965 [3] NCCL INFO comm 0xc3bcba0 rank 5 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfdb804b5f0cc73a2 - Init COMPLETE +ip-26-0-167-9:1080443:1080962 [4] NCCL INFO comm 0xc3975e0 rank 6 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8a730d2677615cb8 - Init COMPLETE +ip-26-0-166-244:1084482:1085000 [4] NCCL INFO comm 0xbfb1330 rank 2 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x8a730d2677615cb8 - Init COMPLETE +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084484:1085002 [6] NCCL INFO comm 0xae5bea0 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8a730d2677615cb8 - Init COMPLETE +ip-26-0-166-244:1084485:1085006 [7] NCCL INFO comm 0xc32e200 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfdb804b5f0cc73a2 - Init COMPLETE +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 02/0 : 3[7] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 03/0 : 3[7] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 02/0 : 5[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 03/0 : 5[3] -> 3[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 02/0 : 5[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 03/0 : 5[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 00/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 01/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 00/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 02/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 01/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 03/0 : 6[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 02/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 02/0 : 5[3] -> 4[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 03/0 : 2[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Channel 03/0 : 5[3] -> 4[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080441:1080960 [2] NCCL INFO comm 0xc3f2e50 rank 5 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8a730d2677615cb8 - Init COMPLETE +ip-26-0-167-9:1080439:1080963 [0] NCCL INFO comm 0xaf579a0 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8a730d2677615cb8 - Init COMPLETE +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 00/0 : 2[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 01/0 : 2[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 00/0 : 4[1] -> 6[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 01/0 : 4[1] -> 6[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 00/0 : 4[1] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 01/0 : 4[1] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 00/0 : 4[1] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 01/0 : 4[1] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 00/0 : 6[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 01/0 : 6[5] -> 4[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 00/0 : 6[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 01/0 : 6[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 02/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 00/0 : 4[1] -> 2[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Channel 03/0 : 1[3] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 01/0 : 4[1] -> 2[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 00/0 : 7[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 01/0 : 7[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 00/0 : 3[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 02/0 : 5[3] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 00/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Channel 03/0 : 5[3] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 01/0 : 3[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 01/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 02/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Channel 03/0 : 6[5] -> 5[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1080967 [1] NCCL INFO comm 0xc4c7d50 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfdb804b5f0cc73a2 - Init COMPLETE +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141901:1142423 [6] NCCL INFO comm 0xb386880 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x303d6fecc3d2249c - Init COMPLETE +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO Connected all trees +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO Connected all trees +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348309:1348865 [4] NCCL INFO comm 0xba52570 rank 2 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x303d6fecc3d2249c - Init COMPLETE +ip-26-0-166-214:1141902:1142428 [7] NCCL INFO comm 0xaed5040 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98423fabe9e87501 - Init COMPLETE +ip-26-0-166-125:1348305:1348863 [0] NCCL INFO comm 0xc186650 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x303d6fecc3d2249c - Init COMPLETE +ip-26-0-166-214:1141899:1142422 [4] NCCL INFO comm 0xbd14d00 rank 6 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x303d6fecc3d2249c - Init COMPLETE +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO Connected all trees +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO Connected all trees +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348306:1348868 [1] NCCL INFO comm 0xb99b0a0 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x98423fabe9e87501 - Init COMPLETE +ip-26-0-166-125:1348307:1348864 [2] NCCL INFO comm 0xbccf9b0 rank 1 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x303d6fecc3d2249c - Init COMPLETE +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO Connected all trees +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO Connected all trees +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348311:1348866 [6] NCCL INFO comm 0xc575280 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x303d6fecc3d2249c - Init COMPLETE +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO Connected all trees +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142426 [3] NCCL INFO comm 0xbe84f60 rank 5 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98423fabe9e87501 - Init COMPLETE +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141900:1142427 [5] NCCL INFO comm 0xbd658c0 rank 6 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x98423fabe9e87501 - Init COMPLETE +ip-26-0-166-125:1348310:1348869 [5] NCCL INFO comm 0xb2a42c0 rank 2 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x98423fabe9e87501 - Init COMPLETE +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348308:1348871 [3] NCCL INFO comm 0xa850e10 rank 1 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98423fabe9e87501 - Init COMPLETE +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO Connected all trees +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141897:1142424 [2] NCCL INFO comm 0xabdb4c0 rank 5 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x303d6fecc3d2249c - Init COMPLETE +ip-26-0-166-214:1141895:1142425 [0] NCCL INFO comm 0xbc032d0 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x303d6fecc3d2249c - Init COMPLETE +ip-26-0-166-125:1348312:1348870 [7] NCCL INFO comm 0xa98d500 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98423fabe9e87501 - Init COMPLETE +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141896:1142429 [1] NCCL INFO comm 0xa9fff80 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x98423fabe9e87501 - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO comm 0xcea2570 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfb94b35786e49ee0 - Init START +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO comm 0x1c340480 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfb94b35786e49ee0 - Init START +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO comm 0xc833440 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x986694d350252cd7 - Init START +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO comm 0x1d27b6a0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x986694d350252cd7 - Init START +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO comm 0xcfcf9b0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x12a02ac7f8006a02 - Init START +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO comm 0x1b45e400 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x12a02ac7f8006a02 - Init START +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO comm 0xba602d0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x42fa1dd6fb304cd4 - Init START +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO comm 0x1c42e980 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x42fa1dd6fb304cd4 - Init START +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Using network Libfabric +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO comm 0xc8cd0e0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x496d52b06e84a03e - Init START +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO comm 0x1bad5d70 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x496d52b06e84a03e - Init START +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO comm 0xb12eb90 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x72f151d6ed148b6b - Init START +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO comm 0x1b98e380 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x72f151d6ed148b6b - Init START +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO comm 0x1b30c7b0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe80e3806f0e8a1ad - Init START +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO comm 0xb6fb310 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe80e3806f0e8a1ad - Init START +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO comm 0x1ca87bb0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x28fbe53bd935b540 - Init START +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO comm 0xb1c4470 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x28fbe53bd935b540 - Init START +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO comm 0xcabaaa0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc798c439a9dcc26c - Init START +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO comm 0x1c1b5ee0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc798c439a9dcc26c - Init START +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO comm 0xcfb1580 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbc1a4d7ab10409b1 - Init START +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO comm 0x1cd02720 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbc1a4d7ab10409b1 - Init START +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Using network Libfabric +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO comm 0xcefc840 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x80ffb5c2900434a - Init START +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO comm 0x1b38a230 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x80ffb5c2900434a - Init START +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO comm 0xcec5ef0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3853daa2bccb0ea0 - Init START +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO comm 0x1cfdde90 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3853daa2bccb0ea0 - Init START +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO comm 0xce37640 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x79699cdb99d54f05 - Init START +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO comm 0x1b56afc0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x79699cdb99d54f05 - Init START +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO comm 0x1d0dd0c0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf5f016c3734cf074 - Init START +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO comm 0xb9654a0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf5f016c3734cf074 - Init START +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO comm 0xb0bfa00 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x20bc8a476788fa35 - Init START +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO comm 0x1d74d9e0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x20bc8a476788fa35 - Init START +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO comm 0x1cc3c960 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe105fbdd3ff7f6b3 - Init START +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO comm 0xcc380e0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe105fbdd3ff7f6b3 - Init START +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Connected all rings +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO Connected all trees +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084481:1085044 [3] NCCL INFO comm 0xb6fb310 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe80e3806f0e8a1ad - Init COMPLETE +ip-26-0-166-244:1084481:1085070 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-166-244:1084481:1085070 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-166-125:1348308:1348918 [3] NCCL INFO comm 0x1b30c7b0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe80e3806f0e8a1ad - Init COMPLETE +ip-26-0-166-125:1348308:1348948 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-166-125:1348308:1348948 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Connected all rings +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO Connected all trees +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Connected all rings +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO Connected all trees +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084480:1085045 [2] NCCL INFO comm 0xb1c4470 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x28fbe53bd935b540 - Init COMPLETE +ip-26-0-166-244:1084480:1085071 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-166-244:1084480:1085071 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348307:1348917 [2] NCCL INFO comm 0x1ca87bb0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x28fbe53bd935b540 - Init COMPLETE +ip-26-0-166-125:1348307:1348949 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-166-125:1348307:1348949 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-166-125:1348309:1348921 [4] NCCL INFO comm 0x1c1b5ee0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc798c439a9dcc26c - Init COMPLETE +ip-26-0-166-125:1348309:1348950 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-125:1348309:1348950 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-244:1084482:1085046 [4] NCCL INFO comm 0xcabaaa0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc798c439a9dcc26c - Init COMPLETE +ip-26-0-166-244:1084482:1085072 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-244:1084482:1085072 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO Connected all trees +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141899:1142473 [4] NCCL INFO comm 0x1c340480 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfb94b35786e49ee0 - Init COMPLETE +ip-26-0-166-214:1141899:1142504 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-214:1141899:1142504 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Connected all rings +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080443:1080994 [4] NCCL INFO comm 0xcea2570 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfb94b35786e49ee0 - Init COMPLETE +ip-26-0-167-9:1080443:1081019 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-167-9:1080443:1081019 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-214:1141900:1142474 [5] NCCL INFO comm 0x1d27b6a0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x986694d350252cd7 - Init COMPLETE +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141900:1142505 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-166-214:1141900:1142505 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-167-9:1080444:1080995 [5] NCCL INFO comm 0xc833440 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x986694d350252cd7 - Init COMPLETE +ip-26-0-167-9:1080444:1081020 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-167-9:1080444:1081020 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Connected all rings +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO Connected all trees +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348310:1348922 [5] NCCL INFO comm 0x1cd02720 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbc1a4d7ab10409b1 - Init COMPLETE +ip-26-0-166-125:1348310:1348951 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-166-125:1348310:1348951 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084483:1085047 [5] NCCL INFO comm 0xcfb1580 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbc1a4d7ab10409b1 - Init COMPLETE +ip-26-0-166-244:1084483:1085073 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-166-244:1084483:1085073 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142486 [3] NCCL INFO comm 0x1cfdde90 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3853daa2bccb0ea0 - Init COMPLETE +ip-26-0-166-214:1141898:1142506 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-166-214:1141898:1142506 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Connected all rings +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Connected all rings +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080442:1081001 [3] NCCL INFO comm 0xcec5ef0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3853daa2bccb0ea0 - Init COMPLETE +ip-26-0-167-9:1080442:1081021 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-167-9:1080442:1081021 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-166-214:1141902:1142481 [7] NCCL INFO comm 0x1b98e380 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x72f151d6ed148b6b - Init COMPLETE +ip-26-0-166-214:1141902:1142507 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-214:1141902:1142507 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-214:1141895:1142478 [0] NCCL INFO comm 0x1c42e980 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x42fa1dd6fb304cd4 - Init COMPLETE +ip-26-0-166-214:1141895:1142508 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-166-214:1141895:1142508 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Connected all rings +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Connected all rings +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO Connected all trees +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Connected all rings +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO Connected all trees +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080446:1080998 [7] NCCL INFO comm 0xb12eb90 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x72f151d6ed148b6b - Init COMPLETE +ip-26-0-167-9:1080446:1081022 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-167-9:1080446:1081022 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Connected all rings +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080439:1080997 [0] NCCL INFO comm 0xba602d0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x42fa1dd6fb304cd4 - Init COMPLETE +ip-26-0-167-9:1080439:1081023 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-167-9:1080439:1081023 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-166-214:1141896:1142477 [1] NCCL INFO comm 0x1b45e400 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x12a02ac7f8006a02 - Init COMPLETE +ip-26-0-166-214:1141896:1142509 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-214:1141896:1142509 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-125:1348311:1348924 [6] NCCL INFO comm 0x1d0dd0c0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf5f016c3734cf074 - Init COMPLETE +ip-26-0-166-125:1348312:1348926 [7] NCCL INFO comm 0x1b56afc0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x79699cdb99d54f05 - Init COMPLETE +ip-26-0-166-125:1348311:1348952 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-166-125:1348311:1348952 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Connected all rings +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO Connected all trees +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348312:1348953 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-125:1348312:1348953 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-214:1141897:1142484 [2] NCCL INFO comm 0x1b38a230 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x80ffb5c2900434a - Init COMPLETE +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141897:1142510 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-166-214:1141897:1142510 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Connected all rings +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Connected all rings +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO Connected all trees +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348306:1348930 [1] NCCL INFO comm 0x1d74d9e0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x20bc8a476788fa35 - Init COMPLETE +ip-26-0-166-125:1348306:1348954 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-125:1348306:1348954 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-244:1084485:1085049 [7] NCCL INFO comm 0xce37640 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x79699cdb99d54f05 - Init COMPLETE +ip-26-0-166-244:1084485:1085074 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-244:1084485:1085074 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-244:1084479:1085051 [1] NCCL INFO comm 0xb0bfa00 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x20bc8a476788fa35 - Init COMPLETE +ip-26-0-166-244:1084479:1085075 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-244:1084479:1085075 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-244:1084484:1085048 [6] NCCL INFO comm 0xb9654a0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf5f016c3734cf074 - Init COMPLETE +ip-26-0-166-244:1084484:1085076 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-166-244:1084484:1085076 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-167-9:1080445:1080999 [6] NCCL INFO comm 0xc8cd0e0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x496d52b06e84a03e - Init COMPLETE +ip-26-0-166-214:1141901:1142482 [6] NCCL INFO comm 0x1bad5d70 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x496d52b06e84a03e - Init COMPLETE +ip-26-0-167-9:1080445:1081024 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-167-9:1080445:1081024 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-167-9:1080440:1080996 [1] NCCL INFO comm 0xcfcf9b0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x12a02ac7f8006a02 - Init COMPLETE +ip-26-0-166-214:1141901:1142511 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-166-214:1141901:1142511 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-167-9:1080440:1081025 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-167-9:1080440:1081025 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-244:1084478:1085050 [0] NCCL INFO comm 0xcc380e0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe105fbdd3ff7f6b3 - Init COMPLETE +ip-26-0-166-244:1084478:1085077 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-166-244:1084478:1085077 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-167-9:1080441:1081000 [2] NCCL INFO comm 0xcefc840 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x80ffb5c2900434a - Init COMPLETE +ip-26-0-167-9:1080441:1081026 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-167-9:1080441:1081026 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-166-125:1348305:1348928 [0] NCCL INFO comm 0x1cc3c960 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe105fbdd3ff7f6b3 - Init COMPLETE +ip-26-0-166-125:1348305:1348955 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-166-125:1348305:1348955 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-166-125:1348308:1349025 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-166-125:1348308:1349025 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-166-125:1348307:1349026 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-166-125:1348307:1349026 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-166-214:1141899:1142580 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-214:1141899:1142580 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-214:1141900:1142581 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-166-214:1141900:1142581 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-166-214:1141901:1142582 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-166-214:1141902:1142583 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-214:1141901:1142582 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-166-214:1141902:1142583 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-125:1348310:1349027 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-166-125:1348310:1349027 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-166-125:1348309:1349028 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-125:1348309:1349028 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-214:1141896:1142585 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-214:1141895:1142586 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-166-214:1141895:1142586 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-166-125:1348305:1349030 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-166-214:1141896:1142585 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-125:1348305:1349030 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-166-125:1348306:1349031 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-125:1348306:1349031 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-125:1348312:1349032 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-125:1348312:1349032 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-125:1348311:1349033 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-166-125:1348311:1349033 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-166-214:1141898:1142587 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-166-214:1141898:1142587 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-166-214:1141897:1142588 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-166-214:1141897:1142588 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-244:1084480:1085162 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-244:1084480:1085162 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-166-244:1084481:1085163 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-166-244:1084481:1085163 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-244:1084482:1085164 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-166-244:1084482:1085164 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-244:1084483:1085165 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-167-9:1080444:1081111 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-166-244:1084483:1085165 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-167-9:1080444:1081111 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-167-9:1080443:1081112 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-167-9:1080443:1081112 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-167-9:1080445:1081113 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-167-9:1080445:1081113 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-167-9:1080446:1081114 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-167-9:1080446:1081114 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-167-9:1080439:1081115 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-167-9:1080439:1081115 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-167-9:1080440:1081116 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-167-9:1080440:1081116 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-244:1084479:1085166 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-166-244:1084479:1085166 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-244:1084478:1085167 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-166-244:1084478:1085167 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-167-9:1080441:1081117 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-167-9:1080441:1081117 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-167-9:1080442:1081118 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-167-9:1080442:1081118 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-166-244:1084485:1085168 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-244:1084485:1085168 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-166-244:1084484:1085169 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-166-244:1084484:1085169 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Using network Libfabric +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Using network Libfabric +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Using network Libfabric +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Using network Libfabric +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO comm 0x1e9f77c0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd144fbe7704a364f - Init START +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO comm 0x1d016410 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd144fbe7704a364f - Init START +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO comm 0x1d15b980 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd144fbe7704a364f - Init START +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO comm 0x202cb930 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd144fbe7704a364f - Init START +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO comm 0x20840440 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x91efe7b95ef5e4fe - Init START +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO comm 0x1e38c6c0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x91efe7b95ef5e4fe - Init START +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO comm 0x1d8433a0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x91efe7b95ef5e4fe - Init START +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO comm 0x20739320 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x91efe7b95ef5e4fe - Init START +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO comm 0x1edd66c0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd9a8894ec0cbd2e2 - Init START +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO comm 0x1e2cd0b0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd9a8894ec0cbd2e2 - Init START +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO comm 0x1ec8bbb0 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd9a8894ec0cbd2e2 - Init START +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO comm 0x1c9957e0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3a180ab34489bb51 - Init START +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO comm 0x1e10f740 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3a180ab34489bb51 - Init START +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO comm 0x1cbf2bb0 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xdd39157f00e6150c - Init START +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO comm 0x20fe5110 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3a180ab34489bb51 - Init START +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO comm 0x204c9590 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd9a8894ec0cbd2e2 - Init START +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO comm 0x1e764040 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xdd39157f00e6150c - Init START +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO comm 0x2077c3e0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xdd39157f00e6150c - Init START +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO comm 0x1e901310 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3a180ab34489bb51 - Init START +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO comm 0x1f72f490 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xdd39157f00e6150c - Init START +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO comm 0x20728310 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x61ec7ef38911bc96 - Init START +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO comm 0x1cae2660 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x61ec7ef38911bc96 - Init START +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO comm 0x1dab0260 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x61ec7ef38911bc96 - Init START +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO comm 0x1f460f90 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x61ec7ef38911bc96 - Init START +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO comm 0x21af3a30 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x95acbad7226e9083 - Init START +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO comm 0x1e65efa0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x95acbad7226e9083 - Init START +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO comm 0x1ca11c90 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x95acbad7226e9083 - Init START +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO comm 0x22131aa0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x95acbad7226e9083 - Init START +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO comm 0x1e90a540 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xaf4b90dfc1e50c7a - Init START +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO comm 0x2040a570 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xaf4b90dfc1e50c7a - Init START +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO comm 0x1d9db1f0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xaf4b90dfc1e50c7a - Init START +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO comm 0x20f47040 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xaf4b90dfc1e50c7a - Init START +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 00/0 : 1[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 01/0 : 1[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 02/0 : 1[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 03/0 : 1[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 00/0 : 3[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 01/0 : 3[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 02/0 : 3[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 03/0 : 3[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 00/0 : 1[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 01/0 : 1[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 02/0 : 1[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 03/0 : 1[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 00/0 : 1[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 01/0 : 1[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 00/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 02/0 : 1[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 03/0 : 1[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 01/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 02/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 03/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 00/0 : 3[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 01/0 : 3[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 02/0 : 3[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 03/0 : 3[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 00/0 : 3[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 01/0 : 3[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 02/0 : 3[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 03/0 : 3[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 00/0 : 1[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 01/0 : 1[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 02/0 : 1[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 03/0 : 1[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 00/0 : 3[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 01/0 : 3[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 02/0 : 3[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 00/0 : 2[0] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 03/0 : 3[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 01/0 : 2[0] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 02/0 : 2[0] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 03/0 : 2[0] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 00/0 : 3[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 01/0 : 3[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 02/0 : 3[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 03/0 : 3[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 00/0 : 2[0] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 01/0 : 2[0] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 02/0 : 2[0] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 03/0 : 2[0] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 00/0 : 3[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 01/0 : 3[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 02/0 : 3[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 03/0 : 3[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 00/0 : 3[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 01/0 : 3[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 02/0 : 3[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 03/0 : 3[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 00/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 01/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 02/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 03/0 : 2[4] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 00/0 : 3[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 01/0 : 3[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 00/0 : 1[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 02/0 : 3[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 03/0 : 3[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 01/0 : 1[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 00/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 02/0 : 1[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 01/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 03/0 : 1[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 00/0 : 1[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 02/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 03/0 : 0[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 01/0 : 1[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 02/0 : 1[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 03/0 : 1[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 00/0 : 3[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 01/0 : 3[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 02/0 : 3[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 03/0 : 3[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 00/0 : 1[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 01/0 : 1[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 02/0 : 1[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 03/0 : 1[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 00/0 : 1[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 01/0 : 1[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 02/0 : 1[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 03/0 : 1[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 00/0 : 3[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 01/0 : 3[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 02/0 : 3[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 03/0 : 3[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Connected all rings +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 00/0 : 2[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 01/0 : 2[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 02/0 : 2[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 03/0 : 2[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Connected all rings +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Connected all rings +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Connected all rings +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 00/0 : 2[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 01/0 : 2[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Connected all rings +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 02/0 : 2[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 03/0 : 2[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Connected all rings +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Connected all rings +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Connected all rings +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Connected all rings +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 00/0 : 3[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 01/0 : 3[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 00/0 : 2[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 01/0 : 2[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 02/0 : 2[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 03/0 : 2[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Connected all rings +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 00/0 : 3[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 00/0 : 2[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Channel 01/0 : 3[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 01/0 : 2[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 02/0 : 2[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 03/0 : 2[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Connected all rings +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Connected all rings +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 00/0 : 3[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 01/0 : 3[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO Connected all trees +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080442:1081130 [3] NCCL INFO comm 0x21af3a30 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x95acbad7226e9083 - Init COMPLETE +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO Connected all trees +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141897:1142604 [2] NCCL INFO comm 0x1ca11c90 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x95acbad7226e9083 - Init COMPLETE +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Connected all rings +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Connected all rings +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 00/0 : 2[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 01/0 : 2[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 00/0 : 3[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Channel 01/0 : 3[1] -> 2[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 02/0 : 2[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 03/0 : 2[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Connected all rings +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Connected all rings +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Connected all rings +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 00/0 : 2[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Connected all rings +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Connected all rings +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 01/0 : 2[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 02/0 : 2[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 03/0 : 2[4] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Connected all rings +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Connected all rings +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Connected all rings +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Connected all rings +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 02/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 00/0 : 3[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Channel 03/0 : 1[5] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 01/0 : 3[5] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Connected all rings +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 00/0 : 3[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 01/0 : 3[1] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Connected all rings +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 00/0 : 2[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 01/0 : 2[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Connected all rings +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 02/0 : 2[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 03/0 : 2[6] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 00/0 : 2[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 01/0 : 2[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 02/0 : 2[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 03/0 : 2[2] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Connected all rings +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 02/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Channel 03/0 : 1[3] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Connected all rings +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 02/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Channel 03/0 : 1[7] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 00/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 01/0 : 3[7] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO Connected all trees +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO Connected all trees +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080441:1081129 [2] NCCL INFO comm 0x22131aa0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x95acbad7226e9083 - Init COMPLETE +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO Connected all trees +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141898:1142606 [3] NCCL INFO comm 0x1e65efa0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x95acbad7226e9083 - Init COMPLETE +ip-26-0-167-9:1080446:1081127 [7] NCCL INFO comm 0x1e9f77c0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd144fbe7704a364f - Init COMPLETE +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO Connected all trees +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO Connected all trees +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO Connected all trees +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO Connected all trees +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080444:1081134 [5] NCCL INFO comm 0x2040a570 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xaf4b90dfc1e50c7a - Init COMPLETE +ip-26-0-166-214:1141901:1142601 [6] NCCL INFO comm 0x1d15b980 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd144fbe7704a364f - Init COMPLETE +ip-26-0-166-214:1141895:1142608 [0] NCCL INFO comm 0x1dab0260 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x61ec7ef38911bc96 - Init COMPLETE +ip-26-0-166-214:1141899:1142609 [4] NCCL INFO comm 0x1d9db1f0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xaf4b90dfc1e50c7a - Init COMPLETE +ip-26-0-167-9:1080440:1081132 [1] NCCL INFO comm 0x20728310 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x61ec7ef38911bc96 - Init COMPLETE +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO Connected all trees +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO Connected all trees +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO Connected all trees +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO Connected all trees +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141900:1142611 [5] NCCL INFO comm 0x1e90a540 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xaf4b90dfc1e50c7a - Init COMPLETE +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO Connected all trees +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080443:1081133 [4] NCCL INFO comm 0x20f47040 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xaf4b90dfc1e50c7a - Init COMPLETE +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO Connected all trees +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080445:1081128 [6] NCCL INFO comm 0x202cb930 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd144fbe7704a364f - Init COMPLETE +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO Connected all trees +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-214:1141902:1142602 [7] NCCL INFO comm 0x1d016410 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd144fbe7704a364f - Init COMPLETE +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-214:1141896:1142610 [1] NCCL INFO comm 0x1cae2660 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x61ec7ef38911bc96 - Init COMPLETE +ip-26-0-166-244:1084483:1085179 [5] NCCL INFO comm 0x20840440 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x91efe7b95ef5e4fe - Init COMPLETE +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO Connected all trees +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-167-9:1080439:1081131 [0] NCCL INFO comm 0x1f460f90 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x61ec7ef38911bc96 - Init COMPLETE +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO Connected all trees +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084481:1085182 [3] NCCL INFO comm 0x20fe5110 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3a180ab34489bb51 - Init COMPLETE +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO Connected all trees +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084485:1085184 [7] NCCL INFO comm 0x2077c3e0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xdd39157f00e6150c - Init COMPLETE +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO Connected all trees +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO Connected all trees +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO Connected all trees +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO Connected all trees +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348309:1349045 [4] NCCL INFO comm 0x1d8433a0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x91efe7b95ef5e4fe - Init COMPLETE +ip-26-0-166-244:1084479:1085180 [1] NCCL INFO comm 0x1ec8bbb0 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd9a8894ec0cbd2e2 - Init COMPLETE +ip-26-0-166-125:1348311:1349054 [6] NCCL INFO comm 0x1e764040 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xdd39157f00e6150c - Init COMPLETE +ip-26-0-166-125:1348305:1349048 [0] NCCL INFO comm 0x1e2cd0b0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd9a8894ec0cbd2e2 - Init COMPLETE +ip-26-0-166-125:1348307:1349050 [2] NCCL INFO comm 0x1e10f740 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3a180ab34489bb51 - Init COMPLETE +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO Connected all trees +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO Connected all trees +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO Connected all trees +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO Connected all trees +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084478:1085181 [0] NCCL INFO comm 0x204c9590 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xd9a8894ec0cbd2e2 - Init COMPLETE +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO Connected all trees +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO Connected all trees +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084482:1085178 [4] NCCL INFO comm 0x20739320 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x91efe7b95ef5e4fe - Init COMPLETE +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO Connected all trees +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-125:1348306:1349051 [1] NCCL INFO comm 0x1edd66c0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xd9a8894ec0cbd2e2 - Init COMPLETE +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO Connected all trees +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-166-244:1084484:1085185 [6] NCCL INFO comm 0x1f72f490 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xdd39157f00e6150c - Init COMPLETE +ip-26-0-166-244:1084480:1085183 [2] NCCL INFO comm 0x1e901310 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3a180ab34489bb51 - Init COMPLETE +ip-26-0-166-125:1348312:1349055 [7] NCCL INFO comm 0x1cbf2bb0 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xdd39157f00e6150c - Init COMPLETE +ip-26-0-166-125:1348310:1349046 [5] NCCL INFO comm 0x1e38c6c0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x91efe7b95ef5e4fe - Init COMPLETE +ip-26-0-166-125:1348308:1349053 [3] NCCL INFO comm 0x1c9957e0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3a180ab34489bb51 - Init COMPLETE +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 223.94 MiB is free. Including non-PyTorch memory, this process has 79.10 GiB memory in use. Of the allocated memory 70.10 GiB is allocated by PyTorch, and 6.89 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/activations.py", line 149, in forward + return nn.functional.silu(input) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/functional.py", line 2072, in silu + return torch._C._nn.silu(input) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 231.94 MiB is free. Including non-PyTorch memory, this process has 79.09 GiB memory in use. Of the allocated memory 69.85 GiB is allocated by PyTorch, and 7.17 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/activations.py", line 149, in forward + return nn.functional.silu(input) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/functional.py", line 2072, in silu + return torch._C._nn.silu(input) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 63.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 69.85 GiB is allocated by PyTorch, and 7.14 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 167.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.62 GiB is allocated by PyTorch, and 6.30 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 514, in backward + total_grad_tensor = torch.empty( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 79.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.35 GiB is allocated by PyTorch, and 6.70 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 223.94 MiB is free. Including non-PyTorch memory, this process has 79.10 GiB memory in use. Of the allocated memory 70.62 GiB is allocated by PyTorch, and 6.52 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[E ProcessGroupGloo.cpp:138] Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +[E ProcessGroupGloo.cpp:138] Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 469, in forward + out = differentiable_reduce_scatter_sum(out, group=group) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 146, in differentiable_reduce_scatter_sum + return DifferentiableReduceScatterSum.apply(tensor, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 119, in forward + dist.reduce_scatter_tensor(sharded_tensor, tensor, group=group, op=dist.ReduceOp.SUM) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 43, in reduce_scatter_tensor + return dist.reduce_scatter_tensor(output=output, input=input, group=group, op=op, async_op=async_op) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3375, in reduce_scatter_tensor + work = group._reduce_scatter_base(output, input, opts) +RuntimeError: Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.125]:18851 +[E ProcessGroupGloo.cpp:2810] [Rank 0]: Rank 1 failed to pass monitoredBarrier in 1200000 ms +[E ProcessGroupGloo.cpp:138] [Rank 0]: Ranks 1 failed to pass monitoredBarrier in 1200000 ms +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 469, in forward + out = differentiable_reduce_scatter_sum(out, group=group) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 146, in differentiable_reduce_scatter_sum + return DifferentiableReduceScatterSum.apply(tensor, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 119, in forward + dist.reduce_scatter_tensor(sharded_tensor, tensor, group=group, op=dist.ReduceOp.SUM) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 43, in reduce_scatter_tensor + return dist.reduce_scatter_tensor(output=output, input=input, group=group, op=op, async_op=async_op) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3375, in reduce_scatter_tensor + work = group._reduce_scatter_base(output, input, opts) +RuntimeError: Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.214]:40087 +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 469, in forward + out = differentiable_reduce_scatter_sum(out, group=group) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 146, in differentiable_reduce_scatter_sum + return DifferentiableReduceScatterSum.apply(tensor, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 119, in forward + dist.reduce_scatter_tensor(sharded_tensor, tensor, group=group, op=dist.ReduceOp.SUM) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 43, in reduce_scatter_tensor + return dist.reduce_scatter_tensor(output=output, input=input, group=group, op=op, async_op=async_op) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3375, in reduce_scatter_tensor + work = group._reduce_scatter_base(output, input, opts) +RuntimeError: [Rank 0]: Ranks 1 failed to pass monitoredBarrier in 1200000 ms +[E ProcessGroupGloo.cpp:138] Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 289, in backward + handle1 = dist.all_gather_into_tensor(unsharded_tensor, tensor, group=group, async_op=True) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 60, in all_gather_into_tensor + return dist.all_gather_into_tensor( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2897, in all_gather_into_tensor + work = group._allgather_base(output_tensor, input_tensor) +RuntimeError: Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.214]:49844 +[E ProcessGroupGloo.cpp:2810] [Rank 0]: Rank 1 failed to pass monitoredBarrier in 1200000 ms +[E ProcessGroupGloo.cpp:138] [Rank 0]: Ranks 1 failed to pass monitoredBarrier in 1200000 ms +[E ProcessGroupGloo.cpp:138] Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +[E ProcessGroupGloo.cpp:138] Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +[E ProcessGroupGloo.cpp:138] Rank 6 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 289, in backward + handle1 = dist.all_gather_into_tensor(unsharded_tensor, tensor, group=group, async_op=True) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 60, in all_gather_into_tensor + return dist.all_gather_into_tensor( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2897, in all_gather_into_tensor + work = group._allgather_base(output_tensor, input_tensor) +RuntimeError: [Rank 0]: Ranks 1 failed to pass monitoredBarrier in 1200000 ms +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 289, in backward + handle1 = dist.all_gather_into_tensor(unsharded_tensor, tensor, group=group, async_op=True) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 60, in all_gather_into_tensor + return dist.all_gather_into_tensor( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2897, in all_gather_into_tensor + work = group._allgather_base(output_tensor, input_tensor) +RuntimeError: Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.125]:31900 +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 529, in training_step + sync_gradients_across_dp( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/data_parallel/utils.py", line 46, in sync_gradients_across_dp + grad_accumulator.sync_gradients_across_dp(dp_pg=dp_pg, reduce_op=reduce_op, **sync_options) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 155, in sync_gradients_across_dp + dist.all_reduce(self._contiguous_fp32_grad_buffer, op=reduce_op, group=dp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + work = group.allreduce([tensor], opts) +RuntimeError: Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:525] Read error [26.0.166.125]:47004: Connection reset by peer +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 529, in training_step + sync_gradients_across_dp( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/data_parallel/utils.py", line 46, in sync_gradients_across_dp + grad_accumulator.sync_gradients_across_dp(dp_pg=dp_pg, reduce_op=reduce_op, **sync_options) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 155, in sync_gradients_across_dp + dist.all_reduce(self._contiguous_fp32_grad_buffer, op=reduce_op, group=dp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + work = group.allreduce([tensor], opts) +RuntimeError: Rank 6 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:525] Read error [26.0.166.125]:47004: Connection reset by peer +[E TraceUtils.h:35] Store is down while updating #1033 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #1033 with key NCCL_0_trace_end +[E TraceUtils.h:35] Store is down while updating #1022 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #1023 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #1235 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #1235 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #1235 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #1235 with key NCCL_1_trace_start +[E ProcessGroupGloo.cpp:138] Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +[E ProcessGroupGloo.cpp:138] Rank 6 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 529, in training_step + sync_gradients_across_dp( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/data_parallel/utils.py", line 46, in sync_gradients_across_dp + grad_accumulator.sync_gradients_across_dp(dp_pg=dp_pg, reduce_op=reduce_op, **sync_options) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 155, in sync_gradients_across_dp + dist.all_reduce(self._contiguous_fp32_grad_buffer, op=reduce_op, group=dp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + work = group.allreduce([tensor], opts) +RuntimeError: Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.125]:15554 +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 529, in training_step + sync_gradients_across_dp( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/data_parallel/utils.py", line 46, in sync_gradients_across_dp + grad_accumulator.sync_gradients_across_dp(dp_pg=dp_pg, reduce_op=reduce_op, **sync_options) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 155, in sync_gradients_across_dp + dist.all_reduce(self._contiguous_fp32_grad_buffer, op=reduce_op, group=dp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + work = group.allreduce([tensor], opts) +RuntimeError: Rank 6 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.125]:47504 +[2024-12-27 18:04:33,207] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 2 (pid: 1348307) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 18:04:33,230] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 3 (pid: 1141898) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 18:04:33,254] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_z2jx3wp_/13803617_m_q8s5es/attempt_0/2/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[0]: + time : 2024-12-27_18:04:04 + host : ip-26-0-166-125.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 1348305) + error_file: /tmp/torchelastic_z2jx3wp_/13803617_m_q8s5es/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 223.94 MiB is free. Including non-PyTorch memory, this process has 79.10 GiB memory in use. Of the allocated memory 70.62 GiB is allocated by PyTorch, and 6.52 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[1]: + time : 2024-12-27_18:04:05 + host : ip-26-0-166-125.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 1348306) + error_file: /tmp/torchelastic_z2jx3wp_/13803617_m_q8s5es/attempt_0/1/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 289, in backward + handle1 = dist.all_gather_into_tensor(unsharded_tensor, tensor, group=group, async_op=True) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 60, in all_gather_into_tensor + return dist.all_gather_into_tensor( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2897, in all_gather_into_tensor + work = group._allgather_base(output_tensor, input_tensor) + RuntimeError: Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: + [../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.125]:31900 + +[3]: + time : 2024-12-27_18:04:04 + host : ip-26-0-166-125.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 1348308) + error_file: /tmp/torchelastic_z2jx3wp_/13803617_m_q8s5es/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 469, in forward + out = differentiable_reduce_scatter_sum(out, group=group) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 146, in differentiable_reduce_scatter_sum + return DifferentiableReduceScatterSum.apply(tensor, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 119, in forward + dist.reduce_scatter_tensor(sharded_tensor, tensor, group=group, op=dist.ReduceOp.SUM) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 43, in reduce_scatter_tensor + return dist.reduce_scatter_tensor(output=output, input=input, group=group, op=op, async_op=async_op) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3375, in reduce_scatter_tensor + work = group._reduce_scatter_base(output, input, opts) + RuntimeError: Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: + [../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.125]:18851 + +[4]: + time : 2024-12-27_18:04:05 + host : ip-26-0-166-125.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 1348309) + error_file: /tmp/torchelastic_z2jx3wp_/13803617_m_q8s5es/attempt_0/4/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 529, in training_step + sync_gradients_across_dp( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/data_parallel/utils.py", line 46, in sync_gradients_across_dp + grad_accumulator.sync_gradients_across_dp(dp_pg=dp_pg, reduce_op=reduce_op, **sync_options) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 155, in sync_gradients_across_dp + dist.all_reduce(self._contiguous_fp32_grad_buffer, op=reduce_op, group=dp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + work = group.allreduce([tensor], opts) + RuntimeError: Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: + [../third_party/gloo/gloo/transport/tcp/pair.cc:525] Read error [26.0.166.125]:47004: Connection reset by peer + +[5]: + time : 2024-12-27_18:04:06 + host : ip-26-0-166-125.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 1348310) + error_file: /tmp/torchelastic_z2jx3wp_/13803617_m_q8s5es/attempt_0/5/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 529, in training_step + sync_gradients_across_dp( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/data_parallel/utils.py", line 46, in sync_gradients_across_dp + grad_accumulator.sync_gradients_across_dp(dp_pg=dp_pg, reduce_op=reduce_op, **sync_options) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 155, in sync_gradients_across_dp + dist.all_reduce(self._contiguous_fp32_grad_buffer, op=reduce_op, group=dp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + work = group.allreduce([tensor], opts) + RuntimeError: Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: + [../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.125]:15554 + +[6]: + time : 2024-12-27_18:04:05 + host : ip-26-0-166-125.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 1348311) + error_file: /tmp/torchelastic_z2jx3wp_/13803617_m_q8s5es/attempt_0/6/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 289, in backward + handle1 = dist.all_gather_into_tensor(unsharded_tensor, tensor, group=group, async_op=True) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 60, in all_gather_into_tensor + return dist.all_gather_into_tensor( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2897, in all_gather_into_tensor + work = group._allgather_base(output_tensor, input_tensor) + RuntimeError: [Rank 0]: Ranks 1 failed to pass monitoredBarrier in 1200000 ms + +[7]: + time : 2024-12-27_18:04:04 + host : ip-26-0-166-125.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 1348312) + error_file: /tmp/torchelastic_z2jx3wp_/13803617_m_q8s5es/attempt_0/7/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 167.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.62 GiB is allocated by PyTorch, and 6.30 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +------------------------------------------------------------ +Root Cause (first observed failure): +[2]: + time : 2024-12-27_18:04:03 + host : ip-26-0-166-125.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 1348307) + error_file: /tmp/torchelastic_z2jx3wp_/13803617_m_q8s5es/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 223.94 MiB is free. Including non-PyTorch memory, this process has 79.10 GiB memory in use. Of the allocated memory 70.10 GiB is allocated by PyTorch, and 6.89 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +============================================================ +[2024-12-27 18:04:33,282] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_rbre78_w/13803617_jblraypp/attempt_0/3/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[0]: + time : 2024-12-27_18:04:04 + host : ip-26-0-166-214.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 1141895) + error_file: /tmp/torchelastic_rbre78_w/13803617_jblraypp/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 514, in backward + total_grad_tensor = torch.empty( + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 79.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.35 GiB is allocated by PyTorch, and 6.70 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[1]: + time : 2024-12-27_18:04:05 + host : ip-26-0-166-214.ec2.internal + rank : 9 (local_rank: 1) + exitcode : 1 (pid: 1141896) + error_file: /tmp/torchelastic_rbre78_w/13803617_jblraypp/attempt_0/1/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 301, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 97, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 289, in backward + handle1 = dist.all_gather_into_tensor(unsharded_tensor, tensor, group=group, async_op=True) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 60, in all_gather_into_tensor + return dist.all_gather_into_tensor( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2897, in all_gather_into_tensor + work = group._allgather_base(output_tensor, input_tensor) + RuntimeError: Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: + [../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.214]:49844 + +[2]: + time : 2024-12-27_18:04:04 + host : ip-26-0-166-214.ec2.internal + rank : 10 (local_rank: 2) + exitcode : 1 (pid: 1141897) + error_file: /tmp/torchelastic_rbre78_w/13803617_jblraypp/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 469, in forward + out = differentiable_reduce_scatter_sum(out, group=group) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 146, in differentiable_reduce_scatter_sum + return DifferentiableReduceScatterSum.apply(tensor, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 119, in forward + dist.reduce_scatter_tensor(sharded_tensor, tensor, group=group, op=dist.ReduceOp.SUM) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 43, in reduce_scatter_tensor + return dist.reduce_scatter_tensor(output=output, input=input, group=group, op=op, async_op=async_op) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3375, in reduce_scatter_tensor + work = group._reduce_scatter_base(output, input, opts) + RuntimeError: [Rank 0]: Ranks 1 failed to pass monitoredBarrier in 1200000 ms + +[4]: + time : 2024-12-27_18:04:05 + host : ip-26-0-166-214.ec2.internal + rank : 12 (local_rank: 4) + exitcode : 1 (pid: 1141899) + error_file: /tmp/torchelastic_rbre78_w/13803617_jblraypp/attempt_0/4/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 529, in training_step + sync_gradients_across_dp( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/data_parallel/utils.py", line 46, in sync_gradients_across_dp + grad_accumulator.sync_gradients_across_dp(dp_pg=dp_pg, reduce_op=reduce_op, **sync_options) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 155, in sync_gradients_across_dp + dist.all_reduce(self._contiguous_fp32_grad_buffer, op=reduce_op, group=dp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + work = group.allreduce([tensor], opts) + RuntimeError: Rank 6 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: + [../third_party/gloo/gloo/transport/tcp/pair.cc:525] Read error [26.0.166.125]:47004: Connection reset by peer + +[5]: + time : 2024-12-27_18:04:06 + host : ip-26-0-166-214.ec2.internal + rank : 13 (local_rank: 5) + exitcode : 1 (pid: 1141900) + error_file: /tmp/torchelastic_rbre78_w/13803617_jblraypp/attempt_0/5/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 529, in training_step + sync_gradients_across_dp( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/data_parallel/utils.py", line 46, in sync_gradients_across_dp + grad_accumulator.sync_gradients_across_dp(dp_pg=dp_pg, reduce_op=reduce_op, **sync_options) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 155, in sync_gradients_across_dp + dist.all_reduce(self._contiguous_fp32_grad_buffer, op=reduce_op, group=dp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2050, in all_reduce + work = group.allreduce([tensor], opts) + RuntimeError: Rank 6 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: + [../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.125]:47504 + +[6]: + time : 2024-12-27_18:04:03 + host : ip-26-0-166-214.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 1141901) + error_file: /tmp/torchelastic_rbre78_w/13803617_jblraypp/attempt_0/6/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/activations.py", line 149, in forward + return nn.functional.silu(input) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/functional.py", line 2072, in silu + return torch._C._nn.silu(input) + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 231.94 MiB is free. Including non-PyTorch memory, this process has 79.09 GiB memory in use. Of the allocated memory 69.85 GiB is allocated by PyTorch, and 7.17 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +[7]: + time : 2024-12-27_18:04:04 + host : ip-26-0-166-214.ec2.internal + rank : 15 (local_rank: 7) + exitcode : 1 (pid: 1141902) + error_file: /tmp/torchelastic_rbre78_w/13803617_jblraypp/attempt_0/7/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 162, in forward + return row_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 593, in row_linear + return _RowLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 469, in forward + out = differentiable_reduce_scatter_sum(out, group=group) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 146, in differentiable_reduce_scatter_sum + return DifferentiableReduceScatterSum.apply(tensor, group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/distributed_differentiable_primitives.py", line 119, in forward + dist.reduce_scatter_tensor(sharded_tensor, tensor, group=group, op=dist.ReduceOp.SUM) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 43, in reduce_scatter_tensor + return dist.reduce_scatter_tensor(output=output, input=input, group=group, op=op, async_op=async_op) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3375, in reduce_scatter_tensor + work = group._reduce_scatter_base(output, input, opts) + RuntimeError: Rank 1 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: + [../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.166.214]:40087 + +------------------------------------------------------------ +Root Cause (first observed failure): +[3]: + time : 2024-12-27_18:04:03 + host : ip-26-0-166-214.ec2.internal + rank : 11 (local_rank: 3) + exitcode : 1 (pid: 1141898) + error_file: /tmp/torchelastic_rbre78_w/13803617_jblraypp/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 161, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 245, in forward + hidden_states = self.down_proj(self.split_silu_mul(merged_states)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 201, in forward + return self.act(gate_states) * up_states + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/nn/activations.py", line 149, in forward + return nn.functional.silu(input) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/functional.py", line 2072, in silu + return torch._C._nn.silu(input) + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 256.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 63.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 69.85 GiB is allocated by PyTorch, and 7.14 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +============================================================ +srun: error: ip-26-0-166-125: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13803617.0 +[2024-12-27 18:04:33,517] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-27 18:04:33,518] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1080439 closing signal SIGTERM +[2024-12-27 18:04:33,518] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1080440 closing signal SIGTERM +[2024-12-27 18:04:33,518] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1080441 closing signal SIGTERM +[2024-12-27 18:04:33,518] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1080442 closing signal SIGTERM +[2024-12-27 18:04:33,518] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1080443 closing signal SIGTERM +[2024-12-27 18:04:33,518] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-27 18:04:33,518] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1080444 closing signal SIGTERM +[2024-12-27 18:04:33,518] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1080445 closing signal SIGTERM +[2024-12-27 18:04:33,519] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1084478 closing signal SIGTERM +[2024-12-27 18:04:33,519] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1084479 closing signal SIGTERM +[2024-12-27 18:04:33,519] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1084480 closing signal SIGTERM +[2024-12-27 18:04:33,519] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1084481 closing signal SIGTERM +[2024-12-27 18:04:33,519] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1084482 closing signal SIGTERM +[2024-12-27 18:04:33,519] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1084483 closing signal SIGTERM +[2024-12-27 18:04:33,519] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1084484 closing signal SIGTERM +[2024-12-27 18:04:33,519] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1084485 closing signal SIGTERM +[2024-12-27 18:04:33,518] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1080446 closing signal SIGTERM +srun: error: ip-26-0-166-214: task 1: Terminated +[2024-12-27 18:04:37,751] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-244.ec2.internal_1084408_0' has failed to send a keep-alive heartbeat to the rendezvous '13803617' due to an error of type RendezvousConnectionError. +[2024-12-27 18:04:38,177] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-9.ec2.internal_1080368_0' has failed to send a keep-alive heartbeat to the rendezvous '13803617' due to an error of type RendezvousConnectionError. +[2024-12-27 18:04:42,753] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-244.ec2.internal_1084408_0' has failed to send a keep-alive heartbeat to the rendezvous '13803617' due to an error of type RendezvousConnectionError. +[2024-12-27 18:04:43,179] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-9.ec2.internal_1080368_0' has failed to send a keep-alive heartbeat to the rendezvous '13803617' due to an error of type RendezvousConnectionError. +[2024-12-27 18:04:47,755] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-244.ec2.internal_1084408_0' has failed to send a keep-alive heartbeat to the rendezvous '13803617' due to an error of type RendezvousConnectionError. +[2024-12-27 18:04:48,180] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-9.ec2.internal_1080368_0' has failed to send a keep-alive heartbeat to the rendezvous '13803617' due to an error of type RendezvousConnectionError. +[2024-12-27 18:04:52,756] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-244.ec2.internal_1084408_0' has failed to send a keep-alive heartbeat to the rendezvous '13803617' due to an error of type RendezvousConnectionError. +[2024-12-27 18:04:53,182] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-9.ec2.internal_1080368_0' has failed to send a keep-alive heartbeat to the rendezvous '13803617' due to an error of type RendezvousConnectionError. +[2024-12-27 18:04:57,758] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-244.ec2.internal_1084408_0' has failed to send a keep-alive heartbeat to the rendezvous '13803617' due to an error of type RendezvousConnectionError. +[2024-12-27 18:04:58,183] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-9.ec2.internal_1080368_0' has failed to send a keep-alive heartbeat to the rendezvous '13803617' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1080368 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1084408 got signal: 15 +srun: error: ip-26-0-167-9: task 3: Exited with exit code 1 +srun: error: ip-26-0-166-244: task 2: Exited with exit code 1 +srun: Force Terminated StepId=13803617.0 diff --git a/logs/13803749-bench_469G_dp2_tp8_pp4_acc32_mbs4_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13803749-bench_469G_dp2_tp8_pp4_acc32_mbs4_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..e73140777dc760b7c1b4f51cefab8e51eafef494 --- /dev/null +++ b/logs/13803749-bench_469G_dp2_tp8_pp4_acc32_mbs4_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,14667 @@ ++ '[' -z 13803749 ']' ++ unset FI_PROVIDER ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-161-[153,178,221],ip-26-0-162-[14,46],ip-26-0-163-[220,226,236]' ++ export 'NODELIST=ip-26-0-161-153 +ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-162-14 +ip-26-0-162-46 +ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236' ++ NODELIST='ip-26-0-161-153 +ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-162-14 +ip-26-0-162-46 +ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-161-[153,178,221],ip-26-0-162-[14,46],ip-26-0-163-[220,226,236]' ++ export MASTER_NODE=ip-26-0-161-153 ++ MASTER_NODE=ip-26-0-161-153 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=8 ++ NNODES=8 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=64 ++ WORLD_SIZE=64 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ export NCCL_P2P_LEVEL=LOC ++ NCCL_P2P_LEVEL=LOC ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-161-153' +Master node: ip-26-0-161-153 ++ echo 'All nodes: ip-26-0-161-153 +ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-162-14 +ip-26-0-162-46 +ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236' +All nodes: ip-26-0-161-153 +ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-162-14 +ip-26-0-162-46 +ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236 ++ echo 'World size: 64' +World size: 64 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=8 --nproc_per_node=8 --rdzv_id=13803749 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-161-153:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_469G_dp2_tp8_pp4_acc32_mbs4_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-27 18:41:39,822] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:41:39,829] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:41:39,835] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:41:39,845] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:41:39,822] torch.distributed.run: [WARNING] +[2024-12-27 18:41:39,822] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,822] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:41:39,822] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,830] torch.distributed.run: [WARNING] +[2024-12-27 18:41:39,830] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,830] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:41:39,830] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,875] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:41:39,835] torch.distributed.run: [WARNING] +[2024-12-27 18:41:39,835] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,835] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:41:39,835] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,881] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:41:39,846] torch.distributed.run: [WARNING] +[2024-12-27 18:41:39,846] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,846] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:41:39,846] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,894] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:41:39,875] torch.distributed.run: [WARNING] +[2024-12-27 18:41:39,875] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,875] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:41:39,875] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,882] torch.distributed.run: [WARNING] +[2024-12-27 18:41:39,882] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,882] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:41:39,882] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,894] torch.distributed.run: [WARNING] +[2024-12-27 18:41:39,894] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:39,894] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:41:39,894] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:40,155] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-27 18:41:40,156] torch.distributed.run: [WARNING] +[2024-12-27 18:41:40,156] torch.distributed.run: [WARNING] ***************************************** +[2024-12-27 18:41:40,156] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-27 18:41:40,156] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-161-153:2456445:2456445 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.161.153<0> +ip-26-0-161-153:2456445:2456445 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-153:2456445:2456445 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-153:2456445:2456445 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-162-46:32948:32948 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32945:32945 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32946:32946 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32947:32947 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-236:883141:883141 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-226:3119069:3119069 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32948:32948 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.162.46<0> +ip-26-0-163-226:3119068:3119068 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32945:32945 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.162.46<0> +ip-26-0-161-153:2456447:2456447 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1600804:1600804 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-178:2582731:2582731 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-236:883144:883144 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32945:32945 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-46:32948:32948 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-178:2582729:2582729 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32948:32948 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-46:32945:32945 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-236:883143:883143 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-236:883145:883145 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-236:883140:883140 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-236:883142:883142 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1600803:1600803 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-178:2582728:2582728 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-236:883146:883146 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-236:883147:883147 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-178:2582730:2582730 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-153:2456449:2456449 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-178:2582726:2582726 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-178:2582733:2582733 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-14:2461730:2461730 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32947:32947 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.162.46<0> +ip-26-0-162-46:32946:32946 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.162.46<0> +ip-26-0-162-14:2461732:2461732 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-178:2582727:2582727 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32947:32947 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-46:32947:32947 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1600809:1600809 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-14:2461731:2461731 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32946:32946 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-46:32946:32946 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-14:2461733:2461733 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-14:2461737:2461737 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-14:2461734:2461734 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-14:2461736:2461736 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-178:2582732:2582732 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-14:2461735:2461735 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32941:32941 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32942:32942 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-226:3119068:3119068 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.163.226<0> +ip-26-0-163-226:3119069:3119069 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.163.226<0> +ip-26-0-162-46:32944:32944 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32943:32943 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-153:2456447:2456447 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.161.153<0> +ip-26-0-163-226:3119072:3119072 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-226:3119069:3119069 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-226:3119068:3119068 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-226:3119068:3119068 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-226:3119069:3119069 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1600810:1600810 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-153:2456449:2456449 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.161.153<0> +ip-26-0-161-153:2456447:2456447 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-153:2456447:2456447 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-153:2456449:2456449 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-153:2456449:2456449 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-153:2456448:2456448 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1600807:1600807 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1600804:1600804 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-226:3119067:3119067 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-226:3119072:3119072 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.163.226<0> +ip-26-0-163-220:1600803:1600803 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-220:1600804:1600804 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1600804:1600804 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-153:2456451:2456451 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32941:32941 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.162.46<0> +ip-26-0-161-153:2456446:2456446 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1600803:1600803 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1600803:1600803 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-226:3119072:3119072 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-226:3119072:3119072 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-153:2456450:2456450 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:30110:30110 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-46:32941:32941 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-46:32941:32941 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-226:3119067:3119067 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.163.226<0> +ip-26-0-163-220:1600805:1600805 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-153:2456452:2456452 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1600806:1600806 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1600808:1600808 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-236:883141:883141 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.163.236<0> +ip-26-0-162-46:32942:32942 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.162.46<0> +ip-26-0-162-46:32944:32944 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.162.46<0> +ip-26-0-162-46:32943:32943 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.162.46<0> +ip-26-0-163-220:1600809:1600809 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-226:3119067:3119067 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:30113:30113 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-226:3119067:3119067 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-46:32943:32943 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-46:32943:32943 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-236:883141:883141 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-153:2456448:2456448 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.161.153<0> +ip-26-0-162-46:32942:32942 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-46:32942:32942 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-236:883141:883141 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-46:32944:32944 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-46:32944:32944 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1600809:1600809 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1600809:1600809 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-236:883144:883144 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.163.236<0> +ip-26-0-161-221:30108:30108 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-236:883140:883140 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.163.236<0> +ip-26-0-163-236:883143:883143 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.163.236<0> +ip-26-0-163-236:883147:883147 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.163.236<0> +ip-26-0-163-236:883146:883146 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.163.236<0> +ip-26-0-163-220:1600810:1600810 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-236:883145:883145 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.163.236<0> +ip-26-0-163-236:883142:883142 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.163.236<0> +ip-26-0-161-221:30112:30112 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-153:2456448:2456448 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-153:2456448:2456448 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:30111:30111 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:30106:30106 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-221:30109:30109 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-236:883144:883144 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-236:883144:883144 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:30107:30107 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1600810:1600810 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1600810:1600810 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-236:883142:883142 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-236:883147:883147 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-236:883143:883143 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-236:883140:883140 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-236:883142:883142 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-236:883147:883147 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-236:883143:883143 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-236:883140:883140 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-236:883146:883146 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-236:883145:883145 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-236:883146:883146 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-236:883145:883145 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1600807:1600807 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-226:3119071:3119071 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-14:2461730:2461730 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.162.14<0> +ip-26-0-162-14:2461732:2461732 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.162.14<0> +ip-26-0-161-153:2456451:2456451 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.161.153<0> +ip-26-0-163-226:3119074:3119074 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-178:2582730:2582730 [4] NCCL INFO Bootstrap : Using enp73s0:26.0.175.184<0> +ip-26-0-161-178:2582731:2582731 [5] NCCL INFO Bootstrap : Using enp73s0:26.0.175.184<0> +ip-26-0-161-178:2582726:2582726 [0] NCCL INFO Bootstrap : Using enp73s0:26.0.175.184<0> +ip-26-0-163-220:1600807:1600807 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1600807:1600807 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-178:2582729:2582729 [3] NCCL INFO Bootstrap : Using enp73s0:26.0.175.184<0> +ip-26-0-161-178:2582733:2582733 [7] NCCL INFO Bootstrap : Using enp73s0:26.0.175.184<0> +ip-26-0-162-14:2461730:2461730 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-14:2461730:2461730 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-178:2582727:2582727 [1] NCCL INFO Bootstrap : Using enp73s0:26.0.175.184<0> +ip-26-0-161-153:2456451:2456451 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-153:2456451:2456451 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-153:2456450:2456450 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.161.153<0> +ip-26-0-161-153:2456452:2456452 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.161.153<0> +ip-26-0-161-153:2456446:2456446 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.161.153<0> +ip-26-0-161-178:2582732:2582732 [6] NCCL INFO Bootstrap : Using enp73s0:26.0.175.184<0> +ip-26-0-161-178:2582728:2582728 [2] NCCL INFO Bootstrap : Using enp73s0:26.0.175.184<0> +ip-26-0-162-14:2461735:2461735 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.162.14<0> +ip-26-0-162-14:2461731:2461731 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.162.14<0> +ip-26-0-162-14:2461736:2461736 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.162.14<0> +ip-26-0-162-14:2461737:2461737 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.162.14<0> +ip-26-0-162-14:2461733:2461733 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.162.14<0> +ip-26-0-162-14:2461734:2461734 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.162.14<0> +ip-26-0-162-14:2461732:2461732 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-14:2461732:2461732 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1600805:1600805 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-220:1600806:1600806 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-220:1600808:1600808 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-161-153:2456450:2456450 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-153:2456450:2456450 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-178:2582726:2582726 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-178:2582730:2582730 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-178:2582730:2582730 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-153:2456452:2456452 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-153:2456452:2456452 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-178:2582726:2582726 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-178:2582731:2582731 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-178:2582729:2582729 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-178:2582731:2582731 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-178:2582729:2582729 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-178:2582732:2582732 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-178:2582732:2582732 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-14:2461735:2461735 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-178:2582728:2582728 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-178:2582728:2582728 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-178:2582733:2582733 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-178:2582727:2582727 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-178:2582733:2582733 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-178:2582727:2582727 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-14:2461735:2461735 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-14:2461736:2461736 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-14:2461737:2461737 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-14:2461733:2461733 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-14:2461736:2461736 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-14:2461737:2461737 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-14:2461733:2461733 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-153:2456446:2456446 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-153:2456446:2456446 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-14:2461731:2461731 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-14:2461731:2461731 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-14:2461734:2461734 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-14:2461734:2461734 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1600805:1600805 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1600805:1600805 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1600806:1600806 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1600806:1600806 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1600808:1600808 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1600808:1600808 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-226:3119073:3119073 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-226:3119071:3119071 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.163.226<0> +ip-26-0-163-226:3119074:3119074 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.163.226<0> +ip-26-0-163-226:3119071:3119071 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-226:3119074:3119074 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-226:3119074:3119074 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-226:3119071:3119071 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-226:3119073:3119073 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.163.226<0> +ip-26-0-163-226:3119070:3119070 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-226:3119073:3119073 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-226:3119073:3119073 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-226:3119070:3119070 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.163.226<0> +ip-26-0-163-226:3119070:3119070 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-226:3119070:3119070 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:30113:30113 [7] NCCL INFO Bootstrap : Using enp72s0:26.0.164.210<0> +ip-26-0-161-221:30110:30110 [4] NCCL INFO Bootstrap : Using enp72s0:26.0.164.210<0> +ip-26-0-161-221:30113:30113 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:30113:30113 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:30106:30106 [0] NCCL INFO Bootstrap : Using enp72s0:26.0.164.210<0> +ip-26-0-161-221:30108:30108 [2] NCCL INFO Bootstrap : Using enp72s0:26.0.164.210<0> +ip-26-0-161-221:30111:30111 [5] NCCL INFO Bootstrap : Using enp72s0:26.0.164.210<0> +ip-26-0-161-221:30112:30112 [6] NCCL INFO Bootstrap : Using enp72s0:26.0.164.210<0> +ip-26-0-161-221:30107:30107 [1] NCCL INFO Bootstrap : Using enp72s0:26.0.164.210<0> +ip-26-0-161-221:30109:30109 [3] NCCL INFO Bootstrap : Using enp72s0:26.0.164.210<0> +ip-26-0-161-221:30106:30106 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:30108:30108 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:30106:30106 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:30107:30107 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:30108:30108 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:30107:30107 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:30111:30111 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:30111:30111 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:30109:30109 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:30112:30112 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:30112:30112 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:30109:30109 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-221:30110:30110 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-221:30110:30110 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Setting provider_filter to efa +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Setting FI_EFA_FORK_SAFE environment variable to 1 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-236:883141:883307 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-236:883141:883307 [1] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/349 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-46:32947:33108 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-46:32947:33108 [6] NCCL INFO Using network Libfabric +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-46:32948:33109 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-46:32948:33109 [7] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Using network Libfabric +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-46:32942:33115 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-46:32942:33115 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:30109:30274 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:30109:30274 [3] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/281 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Using network Libfabric +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-236:883144:883309 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-236:883144:883309 [4] NCCL INFO Using network Libfabric +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-236:883147:883312 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-236:883147:883312 [7] NCCL INFO Using network Libfabric +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-236:883146:883308 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-236:883146:883308 [6] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Using network Libfabric +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-236:883143:883310 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-236:883143:883310 [3] NCCL INFO Using network Libfabric +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-236:883140:883313 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-236:883140:883313 [0] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-236:883145:883311 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-236:883145:883311 [5] NCCL INFO Using network Libfabric +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-236:883142:883314 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-236:883142:883314 [2] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/280 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Using network Libfabric +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-46:32945:33111 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-46:32945:33111 [4] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:30107:30279 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:30107:30279 [1] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:30112:30276 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:30112:30276 [6] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Using network Libfabric +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:30113:30273 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:30113:30273 [7] NCCL INFO Using network Libfabric +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:30111:30275 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:30111:30275 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:30108:30277 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:30108:30277 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:30106:30278 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:30106:30278 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-221:30110:30280 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-221:30110:30280 [4] NCCL INFO Using network Libfabric +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-46:32946:33113 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-46:32946:33113 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Using network Libfabric +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-46:32944:33110 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-46:32944:33110 [3] NCCL INFO Using network Libfabric +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-46:32941:33112 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-46:32941:33112 [0] NCCL INFO Using network Libfabric +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-46:32943:33114 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-46:32943:33114 [2] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO comm 0x87c10c0 rank 31 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO comm 0x9c8b720 rank 30 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-46:32942:33115 [1] NCCL INFO comm 0x8540f50 rank 33 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-46:32941:33112 [0] NCCL INFO comm 0x9207a80 rank 32 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-221:30113:30273 [7] NCCL INFO comm 0x966f990 rank 23 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-221:30112:30276 [6] NCCL INFO comm 0x8ac2940 rank 22 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO comm 0x9031560 rank 28 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO comm 0x8afb230 rank 29 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO comm 0x8c13ec0 rank 26 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO comm 0x954e0e0 rank 27 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO comm 0x8583610 rank 25 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO comm 0x92906b0 rank 24 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-46:32944:33110 [3] NCCL INFO comm 0xa3bfc00 rank 35 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-46:32943:33114 [2] NCCL INFO comm 0x8bed2f0 rank 34 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-46:32945:33111 [4] NCCL INFO comm 0x92ca380 rank 36 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-46:32946:33113 [5] NCCL INFO comm 0x90b6450 rank 37 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-46:32948:33109 [7] NCCL INFO comm 0x9c4fdc0 rank 39 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-162-46:32947:33108 [6] NCCL INFO comm 0x9147db0 rank 38 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO comm 0x9210b10 rank 40 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO comm 0x90af180 rank 47 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO comm 0x86b32c0 rank 45 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO comm 0x9c6f390 rank 44 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO comm 0xa00ede0 rank 41 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO comm 0x9bff270 rank 43 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO comm 0xa2e4320 rank 42 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO comm 0xa088490 rank 46 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO comm 0x8cf9090 rank 49 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO comm 0x96db350 rank 48 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO comm 0x8507b50 rank 51 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO comm 0x8b83dc0 rank 52 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO comm 0x9de7790 rank 53 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO comm 0x9e3c410 rank 55 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO comm 0x882aec0 rank 54 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO comm 0x8d5cb40 rank 50 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-236:883140:883313 [0] NCCL INFO comm 0x8d341a0 rank 56 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-236:883141:883307 [1] NCCL INFO comm 0x8ef9c00 rank 57 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-236:883144:883309 [4] NCCL INFO comm 0x8918b20 rank 60 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-236:883142:883314 [2] NCCL INFO comm 0x907af90 rank 58 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-236:883143:883310 [3] NCCL INFO comm 0x9a3e970 rank 59 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-236:883146:883308 [6] NCCL INFO comm 0x8549930 rank 62 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-236:883147:883312 [7] NCCL INFO comm 0x94a4530 rank 63 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-163-236:883145:883311 [5] NCCL INFO comm 0x8818ba0 rank 61 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO comm 0x8ef90a0 rank 0 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO comm 0x88a4000 rank 1 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO comm 0x88e5f40 rank 2 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO comm 0xa12fb10 rank 3 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-221:30106:30278 [0] NCCL INFO comm 0x8c72ed0 rank 16 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO comm 0x9090b70 rank 15 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO comm 0x90759f0 rank 14 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-221:30111:30275 [5] NCCL INFO comm 0xa3086f0 rank 21 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-221:30110:30280 [4] NCCL INFO comm 0x874c7b0 rank 20 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-221:30108:30277 [2] NCCL INFO comm 0x9729450 rank 18 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO comm 0x8ebc8b0 rank 13 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO comm 0x898b920 rank 12 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO comm 0x8578ab0 rank 11 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO comm 0x8e68050 rank 10 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO comm 0x88e9e80 rank 9 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-221:30109:30274 [3] NCCL INFO comm 0xa1580a0 rank 19 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO comm 0x99ef260 rank 8 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-221:30107:30279 [1] NCCL INFO comm 0x9f77ea0 rank 17 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO comm 0x94c4650 rank 7 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO comm 0xa2d44f0 rank 6 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO comm 0x9d57210 rank 4 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO comm 0x93ec170 rank 5 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init START +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883145:883311 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32948:33109 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32945:33111 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32946:33113 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32942:33115 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32944:33110 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30108:30277 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883140:883313 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/281 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/349 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-236:883147:883312 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883142:883314 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-236:883142:883314 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-46:32946:33113 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-46:32946:33113 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-162-46:32941:33112 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-46:32941:33112 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-236:883145:883311 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-236:883145:883311 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-236:883140:883313 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883140:883313 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883141:883307 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-46:32942:33115 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-236:883141:883307 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-162-46:32942:33115 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-162-46:32945:33111 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-46:32945:33111 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883144:883309 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-236:883144:883309 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883143:883310 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-236:883143:883310 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-46:32948:33109 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-46:32948:33109 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-162-46:32947:33108 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-46:32947:33108 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-46:32943:33114 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-46:32943:33114 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-236:883146:883308 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-236:883146:883308 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-162-46:32944:33110 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-46:32944:33110 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:30112:30276 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:30112:30276 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:30110:30280 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:30110:30280 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:30108:30277 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-221:30108:30277 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30107:30279 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:30107:30279 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30113:30273 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:30113:30273 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30111:30275 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:30111:30275 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30106:30278 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:30106:30278 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30109:30274 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-221:30109:30274 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NCCL_P2P_LEVEL set by environment to LOC +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Trees [0] -1/-1/-1->55->54 [1] -1/-1/-1->55->54 [2] 59/51/-1->55->47 [3] 59/51/-1->55->47 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883147:883312 [7] NCCL INFO Trees [0] -1/-1/-1->63->62 [1] -1/-1/-1->63->62 [2] 31/-1/-1->63->-1 [3] 31/-1/-1->63->-1 +ip-26-0-163-236:883147:883312 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Trees [0] 32/-1/-1->0->-1 [1] 32/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Trees [0] -1/-1/-1->41->42 [1] -1/-1/-1->41->42 [2] 42/40/-1->41->43 [3] 42/40/-1->41->43 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Trees [0] -1/-1/-1->47->46 [1] -1/-1/-1->47->46 [2] 55/39/-1->47->31 [3] 55/39/-1->47->31 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Trees [0] -1/-1/-1->45->46 [1] -1/-1/-1->45->46 [2] 46/44/-1->45->43 [3] 46/44/-1->45->43 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Trees [0] 53/55/-1->54->52 [1] 53/55/-1->54->52 [2] -1/-1/-1->54->53 [3] -1/-1/-1->54->53 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Trees [0] 50/54/-1->52->56 [1] 50/54/-1->52->56 [2] -1/-1/-1->52->53 [3] -1/-1/-1->52->53 +ip-26-0-163-236:883146:883308 [6] NCCL INFO Trees [0] 61/63/-1->62->60 [1] 61/63/-1->62->60 [2] -1/-1/-1->62->61 [3] -1/-1/-1->62->61 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32948:33109 [7] NCCL INFO Trees [0] -1/-1/-1->39->38 [1] -1/-1/-1->39->38 [2] 43/35/-1->39->47 [3] 43/35/-1->39->47 +ip-26-0-162-46:32948:33109 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Trees [0] -1/-1/-1->53->54 [1] -1/-1/-1->53->54 [2] 54/52/-1->53->51 [3] 54/52/-1->53->51 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883146:883308 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883141:883307 [1] NCCL INFO Trees [0] -1/-1/-1->57->58 [1] -1/-1/-1->57->58 [2] 58/56/-1->57->59 [3] 58/56/-1->57->59 +ip-26-0-163-236:883141:883307 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Trees [0] 49/51/-1->50->52 [1] 49/51/-1->50->52 [2] -1/-1/-1->50->49 [3] -1/-1/-1->50->49 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883143:883310 [3] NCCL INFO Trees [0] -1/-1/-1->59->58 [1] -1/-1/-1->59->58 [2] 61/57/-1->59->55 [3] 61/57/-1->59->55 +ip-26-0-163-236:883145:883311 [5] NCCL INFO Trees [0] -1/-1/-1->61->62 [1] -1/-1/-1->61->62 [2] 62/60/-1->61->59 [3] 62/60/-1->61->59 +ip-26-0-163-236:883143:883310 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883145:883311 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32947:33108 [6] NCCL INFO Trees [0] 37/39/-1->38->36 [1] 37/39/-1->38->36 [2] -1/-1/-1->38->37 [3] -1/-1/-1->38->37 +ip-26-0-162-46:32947:33108 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Trees [0] -1/-1/-1->51->50 [1] -1/-1/-1->51->50 [2] 53/49/-1->51->55 [3] 53/49/-1->51->55 +ip-26-0-163-236:883140:883313 [0] NCCL INFO Trees [0] 52/60/-1->56->48 [1] 52/60/-1->56->48 [2] -1/-1/-1->56->57 [3] -1/-1/-1->56->57 +ip-26-0-163-236:883140:883313 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883142:883314 [2] NCCL INFO Trees [0] 57/59/-1->58->60 [1] 57/59/-1->58->60 [2] -1/-1/-1->58->57 [3] -1/-1/-1->58->57 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 47/15/-1->31->63 [3] 47/15/-1->31->63 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Trees [0] 42/46/-1->44->40 [1] 42/46/-1->44->40 [2] -1/-1/-1->44->45 [3] -1/-1/-1->44->45 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Trees [0] 41/43/-1->42->44 [1] 41/43/-1->42->44 [2] -1/-1/-1->42->41 [3] -1/-1/-1->42->41 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32945:33111 [4] NCCL INFO Trees [0] 34/38/-1->36->40 [1] 34/38/-1->36->40 [2] -1/-1/-1->36->37 [3] -1/-1/-1->36->37 +ip-26-0-162-46:32944:33110 [3] NCCL INFO Trees [0] -1/-1/-1->35->34 [1] -1/-1/-1->35->34 [2] 37/33/-1->35->39 [3] 37/33/-1->35->39 +ip-26-0-162-46:32945:33111 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32944:33110 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Trees [0] -1/-1/-1->49->50 [1] -1/-1/-1->49->50 [2] 50/48/-1->49->51 [3] 50/48/-1->49->51 +ip-26-0-163-236:883142:883314 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Trees [0] 45/47/-1->46->44 [1] 45/47/-1->46->44 [2] -1/-1/-1->46->45 [3] -1/-1/-1->46->45 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883144:883309 [4] NCCL INFO Trees [0] 58/62/-1->60->56 [1] 58/62/-1->60->56 [2] -1/-1/-1->60->61 [3] -1/-1/-1->60->61 +ip-26-0-163-236:883144:883309 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-162-46:32942:33115 [1] NCCL INFO Trees [0] -1/-1/-1->33->34 [1] -1/-1/-1->33->34 [2] 34/32/-1->33->35 [3] 34/32/-1->33->35 +ip-26-0-162-46:32942:33115 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Trees [0] 36/44/-1->40->48 [1] 36/44/-1->40->48 [2] -1/-1/-1->40->41 [3] -1/-1/-1->40->41 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Trees [0] 40/56/-1->48->32 [1] 40/56/-1->48->32 [2] -1/-1/-1->48->49 [3] -1/-1/-1->48->49 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32943:33114 [2] NCCL INFO Trees [0] 33/35/-1->34->36 [1] 33/35/-1->34->36 [2] -1/-1/-1->34->33 [3] -1/-1/-1->34->33 +ip-26-0-162-46:32943:33114 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30113:30273 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-161-221:30109:30274 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-161-221:30113:30273 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30109:30274 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30111:30275 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32941:33112 [0] NCCL INFO Trees [0] 16/48/-1->32->0 [1] 16/48/-1->32->0 [2] -1/-1/-1->32->33 [3] -1/-1/-1->32->33 +ip-26-0-162-46:32941:33112 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32946:33113 [5] NCCL INFO Trees [0] -1/-1/-1->37->38 [1] -1/-1/-1->37->38 [2] 38/36/-1->37->35 [3] 38/36/-1->37->35 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32946:33113 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Trees [0] -1/-1/-1->43->42 [1] -1/-1/-1->43->42 [2] 45/41/-1->43->39 [3] 45/41/-1->43->39 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30111:30275 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-161-221:30110:30280 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-161-221:30110:30280 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30112:30276 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-161-221:30107:30279 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-161-221:30112:30276 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30107:30279 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30106:30278 [0] NCCL INFO Trees [0] 8/24/-1->16->32 [1] 8/24/-1->16->32 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-161-221:30106:30278 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30108:30277 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-161-221:30108:30277 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883147:883312 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-46:32948:33109 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-46:32947:33108 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-236:883146:883308 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-236:883141:883307 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:30111:30275 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-236:883143:883310 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-46:32942:33115 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-46:32945:33111 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-236:883145:883311 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-46:32944:33110 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:30109:30274 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-236:883142:883314 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-236:883144:883309 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-236:883140:883313 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-46:32943:33114 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-46:32946:33113 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:30110:30280 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-46:32941:33112 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:30113:30273 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:30107:30279 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:30108:30277 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:30112:30276 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-221:30106:30278 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 00/0 : 38[6] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 01/0 : 38[6] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 02/0 : 38[6] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 03/0 : 38[6] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 00/0 : 39[7] -> 40[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 01/0 : 39[7] -> 40[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 02/0 : 39[7] -> 40[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 03/0 : 39[7] -> 40[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 00/0 : 61[5] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 00/0 : 35[3] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 01/0 : 61[5] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 01/0 : 35[3] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 02/0 : 61[5] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 02/0 : 35[3] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 03/0 : 61[5] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 00/0 : 62[6] -> 63[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 03/0 : 35[3] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 01/0 : 62[6] -> 63[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 00/0 : 32[0] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 02/0 : 62[6] -> 63[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 00/0 : 36[4] -> 37[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 03/0 : 62[6] -> 63[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 01/0 : 32[0] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 01/0 : 36[4] -> 37[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 02/0 : 36[4] -> 37[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 02/0 : 32[0] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 03/0 : 32[0] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 03/0 : 36[4] -> 37[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 00/0 : 33[1] -> 34[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 01/0 : 33[1] -> 34[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 02/0 : 33[1] -> 34[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 03/0 : 33[1] -> 34[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 00/0 : 37[5] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 00/0 : 33[1] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 00/0 : 36[4] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 01/0 : 37[5] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 00/0 : 47[7] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 01/0 : 36[4] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 02/0 : 37[5] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 01/0 : 33[1] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 01/0 : 47[7] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 02/0 : 36[4] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 03/0 : 37[5] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 02/0 : 33[1] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 02/0 : 47[7] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 00/0 : 60[4] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 03/0 : 36[4] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 03/0 : 47[7] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 00/0 : 38[6] -> 39[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 03/0 : 33[1] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 00/0 : 37[5] -> 38[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 01/0 : 38[6] -> 39[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 01/0 : 60[4] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 00/0 : 48[0] -> 49[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 01/0 : 37[5] -> 38[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 00/0 : 34[2] -> 35[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 02/0 : 38[6] -> 39[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 02/0 : 60[4] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 01/0 : 48[0] -> 49[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 02/0 : 37[5] -> 38[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 03/0 : 60[4] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 03/0 : 38[6] -> 39[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 01/0 : 34[2] -> 35[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 03/0 : 37[5] -> 38[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 02/0 : 48[0] -> 49[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 00/0 : 43[3] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 00/0 : 61[5] -> 62[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 00/0 : 31[7] -> 32[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 02/0 : 34[2] -> 35[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 03/0 : 48[0] -> 49[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 00/0 : 40[0] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 00/0 : 55[7] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 03/0 : 34[2] -> 35[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 01/0 : 31[7] -> 32[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 01/0 : 43[3] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 01/0 : 61[5] -> 62[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 00/0 : 31[7] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 02/0 : 61[5] -> 62[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 02/0 : 31[7] -> 32[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 00/0 : 54[6] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 02/0 : 43[3] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 00/0 : 34[2] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 01/0 : 40[0] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 01/0 : 55[7] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 01/0 : 31[7] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 00/0 : 56[0] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 03/0 : 61[5] -> 62[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 03/0 : 31[7] -> 32[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 03/0 : 43[3] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 01/0 : 54[6] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 02/0 : 31[7] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 02/0 : 40[0] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 01/0 : 34[2] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 00/0 : 44[4] -> 45[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 02/0 : 55[7] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 03/0 : 40[0] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 01/0 : 56[0] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 02/0 : 54[6] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 03/0 : 55[7] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 03/0 : 31[7] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 02/0 : 34[2] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 01/0 : 44[4] -> 45[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 00/0 : 41[1] -> 42[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 00/0 : 56[0] -> 57[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 00/0 : 32[0] -> 33[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 00/0 : 53[5] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 02/0 : 56[0] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 02/0 : 44[4] -> 45[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 03/0 : 34[2] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 01/0 : 41[1] -> 42[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 03/0 : 54[6] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 01/0 : 56[0] -> 57[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 03/0 : 44[4] -> 45[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 03/0 : 56[0] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 02/0 : 41[1] -> 42[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 02/0 : 56[0] -> 57[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 01/0 : 32[0] -> 33[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 00/0 : 55[7] -> 56[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 00/0 : 35[3] -> 36[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 01/0 : 53[5] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 00/0 : 49[1] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 00/0 : 45[5] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 00/0 : 57[1] -> 58[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 03/0 : 56[0] -> 57[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 02/0 : 32[0] -> 33[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 00/0 : 39[7] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 00/0 : 57[1] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 01/0 : 35[3] -> 36[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 03/0 : 41[1] -> 42[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 01/0 : 55[7] -> 56[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 00/0 : 44[4] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 02/0 : 53[5] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 01/0 : 57[1] -> 58[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 03/0 : 32[0] -> 33[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 02/0 : 35[3] -> 36[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 01/0 : 45[5] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 01/0 : 39[7] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 00/0 : 51[3] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 02/0 : 55[7] -> 56[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 01/0 : 49[1] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 01/0 : 44[4] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 03/0 : 53[5] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 03/0 : 35[3] -> 36[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 02/0 : 45[5] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 00/0 : 62[6] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 01/0 : 57[1] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 02/0 : 57[1] -> 58[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 02/0 : 39[7] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 02/0 : 49[1] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 03/0 : 55[7] -> 56[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 00/0 : 54[6] -> 55[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 01/0 : 51[3] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 03/0 : 57[1] -> 58[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 02/0 : 57[1] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 02/0 : 44[4] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 03/0 : 45[5] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 03/0 : 39[7] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 03/0 : 49[1] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 01/0 : 54[6] -> 55[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 03/0 : 44[4] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 00/0 : 58[2] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 01/0 : 62[6] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 03/0 : 57[1] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 02/0 : 51[3] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 00/0 : 52[4] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 00/0 : 46[6] -> 47[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 00/0 : 40[0] -> 41[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 00/0 : 45[5] -> 46[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 01/0 : 46[6] -> 47[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 02/0 : 62[6] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 02/0 : 54[6] -> 55[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 00/0 : 50[2] -> 51[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 00/0 : 58[2] -> 59[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 01/0 : 58[2] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 03/0 : 51[3] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 03/0 : 62[6] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 01/0 : 40[0] -> 41[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 01/0 : 45[5] -> 46[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 02/0 : 46[6] -> 47[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 02/0 : 40[0] -> 41[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 01/0 : 52[4] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 00/0 : 50[2] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 01/0 : 58[2] -> 59[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 03/0 : 54[6] -> 55[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 01/0 : 50[2] -> 51[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 00/0 : 52[4] -> 53[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 02/0 : 58[2] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 00/0 : 63[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 02/0 : 45[5] -> 46[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 02/0 : 58[2] -> 59[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 03/0 : 58[2] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 03/0 : 46[6] -> 47[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 02/0 : 52[4] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 03/0 : 40[0] -> 41[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 02/0 : 50[2] -> 51[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 03/0 : 45[5] -> 46[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 01/0 : 52[4] -> 53[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 03/0 : 52[4] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 01/0 : 63[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 03/0 : 58[2] -> 59[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 00/0 : 59[3] -> 60[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 00/0 : 42[2] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 02/0 : 63[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 01/0 : 50[2] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 03/0 : 50[2] -> 51[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 02/0 : 52[4] -> 53[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 00/0 : 53[5] -> 54[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 02/0 : 50[2] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 01/0 : 42[2] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 01/0 : 59[3] -> 60[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 00/0 : 59[3] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 03/0 : 63[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 03/0 : 52[4] -> 53[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 02/0 : 59[3] -> 60[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 01/0 : 53[5] -> 54[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 03/0 : 50[2] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 02/0 : 42[2] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 01/0 : 59[3] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 03/0 : 59[3] -> 60[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 00/0 : 51[3] -> 52[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 02/0 : 59[3] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 03/0 : 42[2] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 02/0 : 53[5] -> 54[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 03/0 : 59[3] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 00/0 : 43[3] -> 44[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 01/0 : 51[3] -> 52[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 00/0 : 60[4] -> 61[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 03/0 : 53[5] -> 54[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 01/0 : 43[3] -> 44[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 02/0 : 51[3] -> 52[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 01/0 : 60[4] -> 61[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 02/0 : 43[3] -> 44[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 03/0 : 51[3] -> 52[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 02/0 : 60[4] -> 61[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 03/0 : 43[3] -> 44[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 03/0 : 60[4] -> 61[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 00/0 : 46[6] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 00/0 : 41[1] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 01/0 : 46[6] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 00/0 : 48[0] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 02/0 : 46[6] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 01/0 : 41[1] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 01/0 : 48[0] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 03/0 : 46[6] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 02/0 : 41[1] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 02/0 : 48[0] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 00/0 : 47[7] -> 48[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 03/0 : 41[1] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 03/0 : 48[0] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 01/0 : 47[7] -> 48[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 00/0 : 49[1] -> 50[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 00/0 : 42[2] -> 43[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 00/0 : 63[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 01/0 : 49[1] -> 50[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 02/0 : 47[7] -> 48[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 02/0 : 49[1] -> 50[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 03/0 : 47[7] -> 48[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 01/0 : 42[2] -> 43[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 03/0 : 49[1] -> 50[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 01/0 : 63[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 02/0 : 42[2] -> 43[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 03/0 : 42[2] -> 43[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 02/0 : 63[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 03/0 : 63[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Connected all rings +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 00/0 : 34[2] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 01/0 : 34[2] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 00/0 : 36[4] -> 38[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 01/0 : 36[4] -> 38[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Connected all rings +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Connected all rings +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Connected all rings +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 02/0 : 41[1] -> 43[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 03/0 : 41[1] -> 43[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Connected all rings +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 02/0 : 33[1] -> 35[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 03/0 : 33[1] -> 35[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Connected all rings +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 00/0 : 34[2] -> 36[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 01/0 : 34[2] -> 36[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Connected all rings +ip-26-0-162-46:32946:33113 [5] NCCL INFO Connected all rings +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 00/0 : 36[4] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 01/0 : 36[4] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 02/0 : 35[3] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 02/0 : 35[3] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 03/0 : 35[3] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 03/0 : 35[3] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 02/0 : 39[7] -> 43[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 03/0 : 39[7] -> 43[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Connected all rings +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 00/0 : 16[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 01/0 : 16[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 00/0 : 32[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 01/0 : 32[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Connected all rings +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 02/0 : 33[1] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 03/0 : 33[1] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 02/0 : 35[3] -> 37[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 03/0 : 35[3] -> 37[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Connected all rings +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 02/0 : 35[3] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 03/0 : 35[3] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 02/0 : 37[5] -> 35[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 02/0 : 35[3] -> 39[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 03/0 : 37[5] -> 35[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 03/0 : 35[3] -> 39[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 02/0 : 39[7] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 03/0 : 39[7] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Connected all rings +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Connected all rings +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Connected all rings +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Connected all rings +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Connected all rings +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 02/0 : 59[3] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 00/0 : 36[4] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 03/0 : 59[3] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 01/0 : 36[4] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 00/0 : 36[4] -> 40[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 00/0 : 38[6] -> 36[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 01/0 : 36[4] -> 40[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 01/0 : 38[6] -> 36[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Connected all rings +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Connected all rings +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Connected all rings +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 02/0 : 31[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 03/0 : 31[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Connected all rings +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 00/0 : 50[2] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 01/0 : 50[2] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 00/0 : 52[4] -> 54[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 01/0 : 52[4] -> 54[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Connected all rings +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Connected all rings +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 02/0 : 57[1] -> 59[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 03/0 : 57[1] -> 59[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Connected all rings +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Connected all rings +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Connected all rings +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 00/0 : 42[2] -> 44[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 01/0 : 42[2] -> 44[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 02/0 : 43[3] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 03/0 : 43[3] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Connected all rings +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Connected all rings +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Connected all rings +ip-26-0-163-236:883143:883310 [3] NCCL INFO Connected all rings +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Connected all rings +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 02/0 : 57[1] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 02/0 : 39[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 03/0 : 57[1] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 00/0 : 50[2] -> 52[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 02/0 : 59[3] -> 61[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 03/0 : 39[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 01/0 : 50[2] -> 52[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 03/0 : 59[3] -> 61[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 02/0 : 47[7] -> 55[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 03/0 : 47[7] -> 55[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 00/0 : 52[4] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 02/0 : 59[3] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 01/0 : 52[4] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 03/0 : 59[3] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 02/0 : 61[5] -> 59[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 02/0 : 55[7] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 03/0 : 61[5] -> 59[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 03/0 : 55[7] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Connected all rings +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Connected all rings +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 00/0 : 40[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 01/0 : 40[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 00/0 : 48[0] -> 56[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 01/0 : 48[0] -> 56[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Connected all rings +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 00/0 : 42[2] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 01/0 : 42[2] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 00/0 : 44[4] -> 46[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 01/0 : 44[4] -> 46[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Connected all rings +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 02/0 : 41[1] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 03/0 : 41[1] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 02/0 : 51[3] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 00/0 : 52[4] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 02/0 : 43[3] -> 45[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 01/0 : 52[4] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 03/0 : 51[3] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 00/0 : 44[4] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 00/0 : 56[0] -> 60[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 03/0 : 43[3] -> 45[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 01/0 : 44[4] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 01/0 : 56[0] -> 60[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 02/0 : 43[3] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 03/0 : 43[3] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 02/0 : 45[5] -> 43[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 02/0 : 39[7] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 03/0 : 45[5] -> 43[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 03/0 : 39[7] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 02/0 : 43[3] -> 39[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 03/0 : 43[3] -> 39[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 02/0 : 39[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 03/0 : 39[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Connected all rings +ip-26-0-161-221:30109:30274 [3] NCCL INFO Connected all rings +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 02/0 : 47[7] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 02/0 : 49[1] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 03/0 : 47[7] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 03/0 : 49[1] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 02/0 : 51[3] -> 53[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 03/0 : 51[3] -> 53[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Connected all rings +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Connected all rings +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 02/0 : 53[5] -> 51[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Connected all rings +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 03/0 : 53[5] -> 51[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 00/0 : 44[4] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 01/0 : 44[4] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 00/0 : 40[0] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 00/0 : 46[6] -> 44[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 01/0 : 40[0] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 01/0 : 46[6] -> 44[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Connected all rings +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 02/0 : 51[3] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 03/0 : 51[3] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Connected all rings +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 02/0 : 55[7] -> 59[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 03/0 : 55[7] -> 59[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 00/0 : 60[4] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 01/0 : 60[4] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 02/0 : 59[3] -> 55[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 03/0 : 59[3] -> 55[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Connected all rings +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 02/0 : 49[1] -> 51[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 03/0 : 49[1] -> 51[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Connected all rings +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 02/0 : 51[3] -> 55[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 02/0 : 51[3] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 03/0 : 51[3] -> 55[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 03/0 : 51[3] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 02/0 : 47[7] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 02/0 : 55[7] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 03/0 : 47[7] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 03/0 : 55[7] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Connected all rings +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 02/0 : 55[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 02/0 : 31[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 03/0 : 55[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 03/0 : 31[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Connected all rings +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 00/0 : 52[4] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 02/0 : 47[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 01/0 : 52[4] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 03/0 : 47[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 00/0 : 52[4] -> 56[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 00/0 : 54[6] -> 52[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Connected all rings +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 01/0 : 52[4] -> 56[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 01/0 : 54[6] -> 52[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 00/0 : 36[4] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 01/0 : 36[4] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 00/0 : 56[0] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 00/0 : 40[0] -> 44[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 01/0 : 40[0] -> 44[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 01/0 : 56[0] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Connected all rings +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 00/0 : 40[0] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 00/0 : 44[4] -> 40[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 01/0 : 40[0] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 00/0 : 40[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 01/0 : 44[4] -> 40[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 01/0 : 40[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 00/0 : 48[0] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 01/0 : 48[0] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Connected all rings +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Connected all rings +ip-26-0-163-236:883144:883309 [4] NCCL INFO Connected all rings +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 00/0 : 58[2] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 02/0 : 31[7] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 01/0 : 58[2] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 03/0 : 31[7] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 00/0 : 60[4] -> 62[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 01/0 : 60[4] -> 62[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 02/0 : 63[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 03/0 : 63[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 00/0 : 62[6] -> 60[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 01/0 : 62[6] -> 60[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Connected all rings +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 00/0 : 58[2] -> 60[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 01/0 : 58[2] -> 60[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 00/0 : 56[0] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 00/0 : 60[4] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 01/0 : 56[0] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 01/0 : 60[4] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 00/0 : 60[4] -> 56[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 00/0 : 48[0] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 01/0 : 60[4] -> 56[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 01/0 : 48[0] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 00/0 : 56[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 00/0 : 32[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 01/0 : 56[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 01/0 : 32[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 00/0 : 48[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 01/0 : 48[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Connected all rings +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Connected all rings +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Connected all rings +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Connected all rings +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Connected all rings +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Connected all rings +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 00/0 : 16[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 01/0 : 16[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 00/0 : 32[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 00/0 : 0[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 01/0 : 32[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 01/0 : 0[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 00/0 : 32[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 01/0 : 32[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Connected all rings +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 00/0 : 32[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 01/0 : 32[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 00/0 : 0[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 01/0 : 0[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 00/0 : 48[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 01/0 : 48[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 00/0 : 32[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 01/0 : 32[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 00/0 : 56[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Connected all rings +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 02/0 : 33[1] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 01/0 : 56[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33112 [0] NCCL INFO Channel 03/0 : 33[1] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 00/0 : 48[0] -> 40[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 01/0 : 48[0] -> 40[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 00/0 : 60[4] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 02/0 : 49[1] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 01/0 : 60[4] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 00/0 : 44[4] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 00/0 : 56[0] -> 52[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 01/0 : 56[0] -> 52[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 01/0 : 44[4] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 00/0 : 40[0] -> 36[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 01/0 : 40[0] -> 36[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 00/0 : 62[6] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 01/0 : 62[6] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Channel 03/0 : 49[1] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 00/0 : 60[4] -> 58[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 00/0 : 46[6] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 02/0 : 57[1] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 00/0 : 54[6] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 01/0 : 46[6] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 01/0 : 60[4] -> 58[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883313 [0] NCCL INFO Channel 03/0 : 57[1] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 01/0 : 54[6] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 02/0 : 41[1] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 00/0 : 52[4] -> 50[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 00/0 : 44[4] -> 42[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Channel 03/0 : 41[1] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 00/0 : 38[6] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 01/0 : 52[4] -> 50[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 00/0 : 63[7] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 01/0 : 38[6] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30278 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 02/0 : 61[5] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 00/0 : 59[3] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 01/0 : 63[7] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 00/0 : 36[4] -> 34[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 01/0 : 44[4] -> 42[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 01/0 : 36[4] -> 34[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883309 [4] NCCL INFO Channel 03/0 : 61[5] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 01/0 : 59[3] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 00/0 : 62[6] -> 61[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 00/0 : 47[7] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 00/0 : 51[3] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 01/0 : 62[6] -> 61[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 00/0 : 58[2] -> 57[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 01/0 : 47[7] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 02/0 : 45[5] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 00/0 : 43[3] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 02/0 : 53[5] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 00/0 : 46[6] -> 45[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 00/0 : 55[7] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 02/0 : 62[6] -> 61[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 01/0 : 58[2] -> 57[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 01/0 : 43[3] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Channel 03/0 : 45[5] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 00/0 : 39[7] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 01/0 : 46[6] -> 45[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 01/0 : 51[3] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Channel 03/0 : 53[5] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883308 [6] NCCL INFO Channel 03/0 : 62[6] -> 61[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 01/0 : 55[7] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 02/0 : 58[2] -> 57[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 00/0 : 42[2] -> 41[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 02/0 : 46[6] -> 45[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 00/0 : 35[3] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 01/0 : 39[7] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 00/0 : 50[2] -> 49[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883314 [2] NCCL INFO Channel 03/0 : 58[2] -> 57[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 00/0 : 54[6] -> 53[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 01/0 : 50[2] -> 49[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 02/0 : 37[5] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 01/0 : 54[6] -> 53[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 02/0 : 50[2] -> 49[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 02/0 : 54[6] -> 53[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 02/0 : 63[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 01/0 : 35[3] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 00/0 : 38[6] -> 37[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33111 [4] NCCL INFO Channel 03/0 : 37[5] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 03/0 : 63[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Channel 03/0 : 50[2] -> 49[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 00/0 : 34[2] -> 33[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 01/0 : 38[6] -> 37[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Channel 03/0 : 54[6] -> 53[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 02/0 : 31[7] -> 63[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 01/0 : 34[2] -> 33[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 02/0 : 38[6] -> 37[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 03/0 : 31[7] -> 63[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 02/0 : 34[2] -> 33[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32947:33108 [6] NCCL INFO Channel 03/0 : 38[6] -> 37[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32943:33114 [2] NCCL INFO Channel 03/0 : 34[2] -> 33[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 02/0 : 47[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 00/0 : 63[7] -> 62[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Channel 01/0 : 63[7] -> 62[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 03/0 : 47[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30280 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30276 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30277 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Channel 03/0 : 46[6] -> 45[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 02/0 : 55[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 03/0 : 55[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 02/0 : 47[7] -> 39[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 03/0 : 47[7] -> 39[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 01/0 : 42[2] -> 41[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 02/0 : 42[2] -> 41[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 00/0 : 47[7] -> 46[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Channel 03/0 : 42[2] -> 41[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 02/0 : 43[3] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Channel 01/0 : 47[7] -> 46[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 03/0 : 43[3] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30273 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 02/0 : 39[7] -> 35[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 03/0 : 39[7] -> 35[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 02/0 : 45[5] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 00/0 : 39[7] -> 38[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 02/0 : 37[5] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 03/0 : 45[5] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 02/0 : 43[3] -> 41[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32948:33109 [7] NCCL INFO Channel 01/0 : 39[7] -> 38[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 03/0 : 43[3] -> 41[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 03/0 : 37[5] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 02/0 : 59[3] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30274 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 02/0 : 35[3] -> 33[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 03/0 : 59[3] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 03/0 : 35[3] -> 33[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 02/0 : 55[7] -> 51[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 00/0 : 46[6] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 03/0 : 55[7] -> 51[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30275 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 00/0 : 43[3] -> 42[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 00/0 : 42[2] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 00/0 : 38[6] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 01/0 : 46[6] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30279 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Channel 01/0 : 43[3] -> 42[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 01/0 : 42[2] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 00/0 : 34[2] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 02/0 : 46[6] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 02/0 : 42[2] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 00/0 : 35[3] -> 34[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 01/0 : 38[6] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 03/0 : 46[6] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 01/0 : 34[2] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33110 [3] NCCL INFO Channel 01/0 : 35[3] -> 34[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 03/0 : 42[2] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 02/0 : 61[5] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 02/0 : 38[6] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 02/0 : 45[5] -> 44[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 00/0 : 55[7] -> 54[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 03/0 : 61[5] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 02/0 : 34[2] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 03/0 : 38[6] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 02/0 : 41[1] -> 40[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 02/0 : 53[5] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Channel 01/0 : 55[7] -> 54[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Channel 03/0 : 45[5] -> 44[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 02/0 : 59[3] -> 57[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 03/0 : 34[2] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 02/0 : 37[5] -> 36[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Channel 03/0 : 41[1] -> 40[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 03/0 : 59[3] -> 57[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 03/0 : 53[5] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 02/0 : 33[1] -> 32[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33113 [5] NCCL INFO Channel 03/0 : 37[5] -> 36[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 02/0 : 51[3] -> 49[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33115 [1] NCCL INFO Channel 03/0 : 33[1] -> 32[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 03/0 : 51[3] -> 49[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 00/0 : 62[6] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 00/0 : 59[3] -> 58[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 01/0 : 62[6] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 00/0 : 58[2] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883143:883310 [3] NCCL INFO Channel 01/0 : 59[3] -> 58[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 02/0 : 62[6] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 00/0 : 54[6] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 01/0 : 58[2] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 03/0 : 62[6] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 00/0 : 50[2] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 00/0 : 51[3] -> 50[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 01/0 : 54[6] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 02/0 : 58[2] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 02/0 : 61[5] -> 60[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Channel 01/0 : 51[3] -> 50[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 03/0 : 58[2] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883311 [5] NCCL INFO Channel 03/0 : 61[5] -> 60[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 02/0 : 54[6] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 01/0 : 50[2] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 02/0 : 57[1] -> 56[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 02/0 : 50[2] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 03/0 : 54[6] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883307 [1] NCCL INFO Channel 03/0 : 57[1] -> 56[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 03/0 : 50[2] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 02/0 : 53[5] -> 52[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 02/0 : 49[1] -> 48[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Channel 03/0 : 53[5] -> 52[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Channel 03/0 : 49[1] -> 48[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883147:883312 [7] NCCL INFO Connected all trees +ip-26-0-163-236:883147:883312 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883147:883312 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883147:883312 [7] NCCL INFO comm 0x94a4530 rank 63 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO Connected all trees +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456445:2456613 [0] NCCL INFO comm 0x8ef90a0 rank 0 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO Connected all trees +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO Connected all trees +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO Connected all trees +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO Connected all trees +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO Connected all trees +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO Connected all trees +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO Connected all trees +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO Connected all trees +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO Connected all trees +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO Connected all trees +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32943:33114 [2] NCCL INFO Connected all trees +ip-26-0-162-46:32943:33114 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32943:33114 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO Connected all trees +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32947:33108 [6] NCCL INFO Connected all trees +ip-26-0-162-46:32947:33108 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32947:33108 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582728:2582896 [2] NCCL INFO comm 0x8e68050 rank 10 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-14:2461736:2461904 [6] NCCL INFO comm 0x9c8b720 rank 30 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-220:1600809:1600975 [6] NCCL INFO comm 0xa088490 rank 46 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-178:2582727:2582892 [1] NCCL INFO comm 0x88e9e80 rank 9 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-221:30112:30276 [6] NCCL INFO Connected all trees +ip-26-0-161-221:30112:30276 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30112:30276 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO Connected all trees +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO Connected all trees +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO Connected all trees +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582732:2582899 [6] NCCL INFO comm 0x90759f0 rank 14 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-226:3119069:3119233 [2] NCCL INFO comm 0x8d5cb40 rank 50 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-178:2582731:2582897 [5] NCCL INFO comm 0x8ebc8b0 rank 13 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-220:1600805:1600979 [2] NCCL INFO comm 0xa2e4320 rank 42 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-14:2461735:2461901 [5] NCCL INFO comm 0x8afb230 rank 29 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-153:2456451:2456616 [6] NCCL INFO comm 0xa2d44f0 rank 6 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-226:3119073:3119239 [6] NCCL INFO comm 0x882aec0 rank 54 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-14:2461732:2461897 [2] NCCL INFO comm 0x8c13ec0 rank 26 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO Connected all trees +ip-26-0-162-46:32943:33114 [2] NCCL INFO comm 0x8bed2f0 rank 34 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32942:33115 [1] NCCL INFO Connected all trees +ip-26-0-162-46:32942:33115 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32942:33115 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32947:33108 [6] NCCL INFO comm 0x9147db0 rank 38 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-153:2456447:2456615 [2] NCCL INFO comm 0x88e5f40 rank 2 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-221:30108:30277 [2] NCCL INFO Connected all trees +ip-26-0-161-221:30108:30277 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30108:30277 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600804:1600976 [1] NCCL INFO comm 0xa00ede0 rank 41 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30112:30276 [6] NCCL INFO comm 0x8ac2940 rank 22 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO Connected all trees +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119072:3119237 [5] NCCL INFO comm 0x9de7790 rank 53 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-46:32946:33113 [5] NCCL INFO Connected all trees +ip-26-0-162-46:32946:33113 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32946:33113 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456446:2456620 [1] NCCL INFO comm 0x88a4000 rank 1 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-153:2456450:2456619 [5] NCCL INFO comm 0x93ec170 rank 5 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-46:32942:33115 [1] NCCL INFO comm 0x8540f50 rank 33 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-14:2461731:2461900 [1] NCCL INFO comm 0x8583610 rank 25 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-221:30108:30277 [2] NCCL INFO comm 0x9729450 rank 18 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-220:1600808:1600981 [5] NCCL INFO comm 0x86b32c0 rank 45 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-226:3119068:3119234 [1] NCCL INFO comm 0x8cf9090 rank 49 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-46:32946:33113 [5] NCCL INFO comm 0x90b6450 rank 37 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-221:30111:30275 [5] NCCL INFO Connected all trees +ip-26-0-161-221:30111:30275 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30111:30275 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30107:30279 [1] NCCL INFO Connected all trees +ip-26-0-161-221:30107:30279 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30107:30279 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30111:30275 [5] NCCL INFO comm 0xa3086f0 rank 21 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-221:30107:30279 [1] NCCL INFO comm 0x9f77ea0 rank 17 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO Connected all trees +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO Connected all trees +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO Connected all trees +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582733:2582895 [7] NCCL INFO comm 0x9090b70 rank 15 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO Connected all trees +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO Connected all trees +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO Connected all trees +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582729:2582898 [3] NCCL INFO comm 0x8578ab0 rank 11 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-14:2461737:2461902 [7] NCCL INFO comm 0x87c10c0 rank 31 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO Connected all trees +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32948:33109 [7] NCCL INFO Connected all trees +ip-26-0-162-46:32948:33109 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32948:33109 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO Connected all trees +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456452:2456618 [7] NCCL INFO comm 0x94c4650 rank 7 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-178:2582730:2582893 [4] NCCL INFO comm 0x898b920 rank 12 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO Connected all trees +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO Connected all trees +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO Connected all trees +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461734:2461903 [4] NCCL INFO comm 0x9031560 rank 28 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32945:33111 [4] NCCL INFO Connected all trees +ip-26-0-162-46:32945:33111 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32945:33111 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30109:30274 [3] NCCL INFO Connected all trees +ip-26-0-161-221:30109:30274 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30109:30274 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO Connected all trees +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32948:33109 [7] NCCL INFO comm 0x9c4fdc0 rank 39 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO Connected all trees +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO Connected all trees +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30110:30280 [4] NCCL INFO Connected all trees +ip-26-0-161-221:30110:30280 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30110:30280 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600806:1600980 [3] NCCL INFO comm 0x9bff270 rank 43 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-221:30113:30273 [7] NCCL INFO Connected all trees +ip-26-0-161-221:30113:30273 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30113:30273 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32944:33110 [3] NCCL INFO Connected all trees +ip-26-0-162-46:32944:33110 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32944:33110 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582726:2582894 [0] NCCL INFO comm 0x99ef260 rank 8 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-220:1600810:1600978 [7] NCCL INFO comm 0x90af180 rank 47 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-221:30106:30278 [0] NCCL INFO Connected all trees +ip-26-0-161-221:30106:30278 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30106:30278 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461733:2461899 [3] NCCL INFO comm 0x954e0e0 rank 27 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-226:3119074:3119238 [7] NCCL INFO comm 0x9e3c410 rank 55 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-14:2461730:2461898 [0] NCCL INFO comm 0x92906b0 rank 24 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-220:1600803:1600974 [0] NCCL INFO comm 0x9210b10 rank 40 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-46:32941:33112 [0] NCCL INFO Connected all trees +ip-26-0-162-46:32941:33112 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32941:33112 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456449:2456614 [4] NCCL INFO comm 0x9d57210 rank 4 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-46:32945:33111 [4] NCCL INFO comm 0x92ca380 rank 36 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-221:30109:30274 [3] NCCL INFO comm 0xa1580a0 rank 19 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-220:1600807:1600977 [4] NCCL INFO comm 0x9c6f390 rank 44 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO Connected all trees +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119070:3119235 [3] NCCL INFO comm 0x8507b50 rank 51 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-226:3119071:3119236 [4] NCCL INFO comm 0x8b83dc0 rank 52 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-226:3119067:3119240 [0] NCCL INFO comm 0x96db350 rank 48 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-221:30110:30280 [4] NCCL INFO comm 0x874c7b0 rank 20 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-46:32944:33110 [3] NCCL INFO comm 0xa3bfc00 rank 35 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-221:30113:30273 [7] NCCL INFO comm 0x966f990 rank 23 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-221:30106:30278 [0] NCCL INFO comm 0x8c72ed0 rank 16 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-162-46:32941:33112 [0] NCCL INFO comm 0x9207a80 rank 32 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-153:2456448:2456617 [3] NCCL INFO comm 0xa12fb10 rank 3 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-236:883142:883314 [2] NCCL INFO Connected all trees +ip-26-0-163-236:883142:883314 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883142:883314 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883142:883314 [2] NCCL INFO comm 0x907af90 rank 58 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-236:883146:883308 [6] NCCL INFO Connected all trees +ip-26-0-163-236:883146:883308 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883146:883308 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883145:883311 [5] NCCL INFO Connected all trees +ip-26-0-163-236:883145:883311 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883145:883311 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883141:883307 [1] NCCL INFO Connected all trees +ip-26-0-163-236:883141:883307 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883141:883307 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883146:883308 [6] NCCL INFO comm 0x8549930 rank 62 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-236:883145:883311 [5] NCCL INFO comm 0x8818ba0 rank 61 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-236:883141:883307 [1] NCCL INFO comm 0x8ef9c00 rank 57 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-236:883143:883310 [3] NCCL INFO Connected all trees +ip-26-0-163-236:883143:883310 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883143:883310 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883143:883310 [3] NCCL INFO comm 0x9a3e970 rank 59 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-236:883140:883313 [0] NCCL INFO Connected all trees +ip-26-0-163-236:883140:883313 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883140:883313 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883144:883309 [4] NCCL INFO Connected all trees +ip-26-0-163-236:883144:883309 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883144:883309 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883140:883313 [0] NCCL INFO comm 0x8d341a0 rank 56 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-163-236:883144:883309 [4] NCCL INFO comm 0x8918b20 rank 60 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x392a5880c0b1f7a4 - Init COMPLETE +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Using network Libfabric +ip-26-0-163-236:883141:883507 [1] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Using network Libfabric +ip-26-0-162-46:32948:33309 [7] NCCL INFO Using network Libfabric +ip-26-0-162-46:32946:33310 [5] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Using network Libfabric +ip-26-0-162-46:32942:33312 [1] NCCL INFO Using network Libfabric +ip-26-0-162-46:32947:33313 [6] NCCL INFO Using network Libfabric +ip-26-0-162-46:32944:33314 [3] NCCL INFO Using network Libfabric +ip-26-0-163-236:883145:883509 [5] NCCL INFO Using network Libfabric +ip-26-0-163-236:883144:883510 [4] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Using network Libfabric +ip-26-0-163-236:883143:883511 [3] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Using network Libfabric +ip-26-0-163-236:883142:883508 [2] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Using network Libfabric +ip-26-0-163-236:883147:883513 [7] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Using network Libfabric +ip-26-0-163-236:883140:883506 [0] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:30111:30485 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:30109:30490 [3] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Using network Libfabric +ip-26-0-162-46:32945:33308 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:30110:30488 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:30108:30491 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:30112:30486 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:30106:30489 [0] NCCL INFO Using network Libfabric +ip-26-0-162-46:32941:33307 [0] NCCL INFO Using network Libfabric +ip-26-0-163-236:883146:883512 [6] NCCL INFO Using network Libfabric +ip-26-0-162-46:32943:33311 [2] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Using network Libfabric +ip-26-0-161-221:30107:30492 [1] NCCL INFO Using network Libfabric +ip-26-0-161-221:30113:30487 [7] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO comm 0x92cea80 rank 15 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-221:30106:30489 [0] NCCL INFO comm 0x8eb32c0 rank 16 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-221:30108:30491 [2] NCCL INFO comm 0x996a8e0 rank 18 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-221:30107:30492 [1] NCCL INFO comm 0xa1b8af0 rank 17 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO comm 0x9702ac0 rank 7 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO comm 0xa5138b0 rank 6 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO comm 0x9f96a40 rank 4 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO comm 0x92b4e20 rank 14 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO comm 0x90fd0c0 rank 13 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO comm 0x962bdf0 rank 5 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO comm 0x8bcb770 rank 12 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO comm 0xa36f7b0 rank 3 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO comm 0x87b7ba0 rank 11 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO comm 0xa07bf70 rank 55 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-236:883147:883513 [7] NCCL INFO comm 0x96e3880 rank 63 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-236:883144:883510 [4] NCCL INFO comm 0x8b57490 rank 60 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-236:883146:883512 [6] NCCL INFO comm 0x8789af0 rank 62 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-236:883145:883509 [5] NCCL INFO comm 0x8a56780 rank 61 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-236:883143:883511 [3] NCCL INFO comm 0x9c7e370 rank 59 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO comm 0x8b24900 rank 2 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO comm 0x8b297f0 rank 9 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO comm 0x9c2ebc0 rank 8 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO comm 0x90a72e0 rank 10 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO comm 0x8ae4240 rank 1 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO comm 0x9e3f460 rank 43 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO comm 0x92efbb0 rank 47 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO comm 0xa2c8e00 rank 46 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO comm 0x8a6a4b0 rank 54 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO comm 0x9146c80 rank 0 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO comm 0xa025180 rank 53 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO comm 0x8dc8a40 rank 52 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-236:883141:883507 [1] NCCL INFO comm 0x913ad70 rank 57 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-236:883142:883508 [2] NCCL INFO comm 0x92b9780 rank 58 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO comm 0x8746f20 rank 51 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO comm 0x8f3a9f0 rank 49 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO comm 0x8f9ba40 rank 50 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-236:883140:883506 [0] NCCL INFO comm 0x8f73830 rank 56 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-46:32948:33309 [7] NCCL INFO comm 0x9e8fff0 rank 39 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-46:32947:33313 [6] NCCL INFO comm 0x9387380 rank 38 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO comm 0x88f3960 rank 45 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO comm 0xa523e50 rank 42 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO comm 0x9eae400 rank 44 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO comm 0x99200c0 rank 48 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO comm 0x8a01b30 rank 31 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-46:32946:33310 [5] NCCL INFO comm 0x92f53c0 rank 37 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO comm 0xa24efc0 rank 41 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-221:30111:30485 [5] NCCL INFO comm 0xa547fc0 rank 21 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-46:32945:33308 [4] NCCL INFO comm 0x950aa70 rank 36 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init START +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO comm 0x9450990 rank 40 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-46:32944:33314 [3] NCCL INFO comm 0xa5fe960 rank 35 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-46:32943:33311 [2] NCCL INFO comm 0x8e2a5d0 rank 34 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-46:32942:33312 [1] NCCL INFO comm 0x87800f0 rank 33 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO comm 0x9eca7c0 rank 30 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO comm 0x8d3a680 rank 29 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO comm 0x9271080 rank 28 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-46:32941:33307 [0] NCCL INFO comm 0x9447340 rank 32 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-221:30112:30486 [6] NCCL INFO comm 0x8d03370 rank 22 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-221:30113:30487 [7] NCCL INFO comm 0x98b0870 rank 23 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO comm 0x978e300 rank 27 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO comm 0x87c3ac0 rank 25 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO comm 0x8e52fc0 rank 26 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-221:30110:30488 [4] NCCL INFO comm 0x898ca30 rank 20 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-221:30109:30490 [3] NCCL INFO comm 0xa39a640 rank 19 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init START +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO comm 0x94d23b0 rank 24 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init START +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30106:30489 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30107:30492 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30108:30491 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883146:883512 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883147:883513 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32948:33309 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30110:30488 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/281 +ip-26-0-161-221:30109:30490 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883145:883509 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/349 +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32947:33313 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883144:883510 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32942:33312 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32946:33310 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883143:883511 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32941:33307 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32944:33314 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30111:30485 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883141:883507 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883142:883508 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32945:33308 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30112:30486 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30113:30487 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883140:883506 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32943:33311 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883143:883511 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-236:883146:883512 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-162-46:32941:33307 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-236:883140:883506 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883141:883507 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-236:883144:883510 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883147:883513 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-236:883142:883508 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-236:883145:883509 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-221:30111:30485 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-221:30108:30491 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-221:30109:30490 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30107:30492 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-162-46:32948:33309 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:30106:30489 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-46:32942:33312 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-162-46:32943:33311 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-221:30112:30486 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-46:32946:33310 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:30110:30488 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-46:32945:33308 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:30113:30487 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-162-46:32947:33313 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-46:32944:33314 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-46:32944:33314 [3] NCCL INFO Trees [0] -1/-1/-1->35->34 [1] -1/-1/-1->35->34 [2] 37/33/-1->35->39 [3] 37/33/-1->35->39 +ip-26-0-162-46:32944:33314 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32945:33308 [4] NCCL INFO Trees [0] 34/38/-1->36->40 [1] 34/38/-1->36->40 [2] -1/-1/-1->36->37 [3] -1/-1/-1->36->37 +ip-26-0-162-46:32945:33308 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32947:33313 [6] NCCL INFO Trees [0] 37/39/-1->38->36 [1] 37/39/-1->38->36 [2] -1/-1/-1->38->37 [3] -1/-1/-1->38->37 +ip-26-0-162-46:32947:33313 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32946:33310 [5] NCCL INFO Trees [0] -1/-1/-1->37->38 [1] -1/-1/-1->37->38 [2] 38/36/-1->37->35 [3] 38/36/-1->37->35 +ip-26-0-162-46:32948:33309 [7] NCCL INFO Trees [0] -1/-1/-1->39->38 [1] -1/-1/-1->39->38 [2] 43/35/-1->39->47 [3] 43/35/-1->39->47 +ip-26-0-162-46:32946:33310 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32948:33309 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Trees [0] 36/44/-1->40->48 [1] 36/44/-1->40->48 [2] -1/-1/-1->40->41 [3] -1/-1/-1->40->41 +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Trees [0] -1/-1/-1->41->42 [1] -1/-1/-1->41->42 [2] 42/40/-1->41->43 [3] 42/40/-1->41->43 +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Trees [0] -1/-1/-1->55->54 [1] -1/-1/-1->55->54 [2] 59/51/-1->55->47 [3] 59/51/-1->55->47 +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Trees [0] 41/43/-1->42->44 [1] 41/43/-1->42->44 [2] -1/-1/-1->42->41 [3] -1/-1/-1->42->41 +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Trees [0] -1/-1/-1->43->42 [1] -1/-1/-1->43->42 [2] 45/41/-1->43->39 [3] 45/41/-1->43->39 +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Trees [0] -1/-1/-1->45->46 [1] -1/-1/-1->45->46 [2] 46/44/-1->45->43 [3] 46/44/-1->45->43 +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Trees [0] -1/-1/-1->47->46 [1] -1/-1/-1->47->46 [2] 55/39/-1->47->31 [3] 55/39/-1->47->31 +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Trees [0] 45/47/-1->46->44 [1] 45/47/-1->46->44 [2] -1/-1/-1->46->45 [3] -1/-1/-1->46->45 +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Trees [0] 53/55/-1->54->52 [1] 53/55/-1->54->52 [2] -1/-1/-1->54->53 [3] -1/-1/-1->54->53 +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Trees [0] -1/-1/-1->53->54 [1] -1/-1/-1->53->54 [2] 54/52/-1->53->51 [3] 54/52/-1->53->51 +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Trees [0] 50/54/-1->52->56 [1] 50/54/-1->52->56 [2] -1/-1/-1->52->53 [3] -1/-1/-1->52->53 +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Trees [0] -1/-1/-1->51->50 [1] -1/-1/-1->51->50 [2] 53/49/-1->51->55 [3] 53/49/-1->51->55 +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Trees [0] 42/46/-1->44->40 [1] 42/46/-1->44->40 [2] -1/-1/-1->44->45 [3] -1/-1/-1->44->45 +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883141:883507 [1] NCCL INFO Trees [0] -1/-1/-1->57->58 [1] -1/-1/-1->57->58 [2] 58/56/-1->57->59 [3] 58/56/-1->57->59 +ip-26-0-163-236:883141:883507 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Trees [0] 13/15/-1->14->12 [1] 13/15/-1->14->12 [2] -1/-1/-1->14->13 [3] -1/-1/-1->14->13 +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Trees [0] -1/-1/-1->13->14 [1] -1/-1/-1->13->14 [2] 14/12/-1->13->11 [3] 14/12/-1->13->11 +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Trees [0] 10/14/-1->12->8 [1] 10/14/-1->12->8 [2] -1/-1/-1->12->13 [3] -1/-1/-1->12->13 +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] -1/-1/-1->15->14 [2] 23/7/-1->15->31 [3] 23/7/-1->15->31 +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Trees [0] -1/-1/-1->29->30 [1] -1/-1/-1->29->30 [2] 30/28/-1->29->27 [3] 30/28/-1->29->27 +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Trees [0] 29/31/-1->30->28 [1] 29/31/-1->30->28 [2] -1/-1/-1->30->29 [3] -1/-1/-1->30->29 +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Trees [0] 49/51/-1->50->52 [1] 49/51/-1->50->52 [2] -1/-1/-1->50->49 [3] -1/-1/-1->50->49 +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Trees [0] 40/56/-1->48->32 [1] 40/56/-1->48->32 [2] -1/-1/-1->48->49 [3] -1/-1/-1->48->49 +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Trees [0] -1/-1/-1->49->50 [1] -1/-1/-1->49->50 [2] 50/48/-1->49->51 [3] 50/48/-1->49->51 +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30107:30492 [1] NCCL INFO Trees [0] -1/-1/-1->17->18 [1] -1/-1/-1->17->18 [2] 18/16/-1->17->19 [3] 18/16/-1->17->19 +ip-26-0-161-221:30107:30492 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883142:883508 [2] NCCL INFO Trees [0] 57/59/-1->58->60 [1] 57/59/-1->58->60 [2] -1/-1/-1->58->57 [3] -1/-1/-1->58->57 +ip-26-0-163-236:883147:883513 [7] NCCL INFO Trees [0] -1/-1/-1->63->62 [1] -1/-1/-1->63->62 [2] 31/-1/-1->63->-1 [3] 31/-1/-1->63->-1 +ip-26-0-163-236:883142:883508 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883147:883513 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32941:33307 [0] NCCL INFO Trees [0] 16/48/-1->32->0 [1] 16/48/-1->32->0 [2] -1/-1/-1->32->33 [3] -1/-1/-1->32->33 +ip-26-0-162-46:32942:33312 [1] NCCL INFO Trees [0] -1/-1/-1->33->34 [1] -1/-1/-1->33->34 [2] 34/32/-1->33->35 [3] 34/32/-1->33->35 +ip-26-0-162-46:32943:33311 [2] NCCL INFO Trees [0] 33/35/-1->34->36 [1] 33/35/-1->34->36 [2] -1/-1/-1->34->33 [3] -1/-1/-1->34->33 +ip-26-0-162-46:32941:33307 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32942:33312 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32943:33311 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883145:883509 [5] NCCL INFO Trees [0] -1/-1/-1->61->62 [1] -1/-1/-1->61->62 [2] 62/60/-1->61->59 [3] 62/60/-1->61->59 +ip-26-0-163-236:883146:883512 [6] NCCL INFO Trees [0] 61/63/-1->62->60 [1] 61/63/-1->62->60 [2] -1/-1/-1->62->61 [3] -1/-1/-1->62->61 +ip-26-0-163-236:883145:883509 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883143:883511 [3] NCCL INFO Trees [0] -1/-1/-1->59->58 [1] -1/-1/-1->59->58 [2] 61/57/-1->59->55 [3] 61/57/-1->59->55 +ip-26-0-163-236:883146:883512 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Trees [0] -1/-1/-1->11->10 [1] -1/-1/-1->11->10 [2] 13/9/-1->11->7 [3] 13/9/-1->11->7 +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Trees [0] 9/11/-1->10->12 [1] 9/11/-1->10->12 [2] -1/-1/-1->10->9 [3] -1/-1/-1->10->9 +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Trees [0] 4/12/-1->8->16 [1] 4/12/-1->8->16 [2] -1/-1/-1->8->9 [3] -1/-1/-1->8->9 +ip-26-0-163-236:883143:883511 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Trees [0] 26/30/-1->28->24 [1] 26/30/-1->28->24 [2] -1/-1/-1->28->29 [3] -1/-1/-1->28->29 +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] -1/-1/-1->31->30 [2] 47/15/-1->31->63 [3] 47/15/-1->31->63 +ip-26-0-163-236:883144:883510 [4] NCCL INFO Trees [0] 58/62/-1->60->56 [1] 58/62/-1->60->56 [2] -1/-1/-1->60->61 [3] -1/-1/-1->60->61 +ip-26-0-163-236:883140:883506 [0] NCCL INFO Trees [0] 52/60/-1->56->48 [1] 52/60/-1->56->48 [2] -1/-1/-1->56->57 [3] -1/-1/-1->56->57 +ip-26-0-163-236:883144:883510 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883140:883506 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 11/3/-1->7->15 [3] 11/3/-1->7->15 +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Trees [0] -1/-1/-1->9->10 [1] -1/-1/-1->9->10 [2] 10/8/-1->9->11 [3] 10/8/-1->9->11 +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30113:30487 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] -1/-1/-1->23->22 [2] 27/19/-1->23->15 [3] 27/19/-1->23->15 +ip-26-0-161-221:30113:30487 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Trees [0] 25/27/-1->26->28 [1] 25/27/-1->26->28 [2] -1/-1/-1->26->25 [3] -1/-1/-1->26->25 +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Trees [0] -1/-1/-1->27->26 [1] -1/-1/-1->27->26 [2] 29/25/-1->27->23 [3] 29/25/-1->27->23 +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30109:30490 [3] NCCL INFO Trees [0] -1/-1/-1->19->18 [1] -1/-1/-1->19->18 [2] 21/17/-1->19->23 [3] 21/17/-1->19->23 +ip-26-0-161-221:30109:30490 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30106:30489 [0] NCCL INFO Trees [0] 8/24/-1->16->32 [1] 8/24/-1->16->32 [2] -1/-1/-1->16->17 [3] -1/-1/-1->16->17 +ip-26-0-161-221:30111:30485 [5] NCCL INFO Trees [0] -1/-1/-1->21->22 [1] -1/-1/-1->21->22 [2] 22/20/-1->21->19 [3] 22/20/-1->21->19 +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Trees [0] 20/28/-1->24->16 [1] 20/28/-1->24->16 [2] -1/-1/-1->24->25 [3] -1/-1/-1->24->25 +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Trees [0] -1/-1/-1->25->26 [1] -1/-1/-1->25->26 [2] 26/24/-1->25->27 [3] 26/24/-1->25->27 +ip-26-0-161-221:30110:30488 [4] NCCL INFO Trees [0] 18/22/-1->20->24 [1] 18/22/-1->20->24 [2] -1/-1/-1->20->21 [3] -1/-1/-1->20->21 +ip-26-0-161-221:30106:30489 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30111:30485 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30112:30486 [6] NCCL INFO Trees [0] 21/23/-1->22->20 [1] 21/23/-1->22->20 [2] -1/-1/-1->22->21 [3] -1/-1/-1->22->21 +ip-26-0-161-221:30110:30488 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30112:30486 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Trees [0] 32/-1/-1->0->-1 [1] 32/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Trees [0] 2/6/-1->4->8 [1] 2/6/-1->4->8 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30108:30491 [2] NCCL INFO Trees [0] 17/19/-1->18->20 [1] 17/19/-1->18->20 [2] -1/-1/-1->18->17 [3] -1/-1/-1->18->17 +ip-26-0-161-221:30108:30491 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 00/0 : 33[1] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 00/0 : 35[3] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 01/0 : 33[1] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 02/0 : 33[1] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 00/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 01/0 : 35[3] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 03/0 : 33[1] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 02/0 : 35[3] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 01/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 00/0 : 34[2] -> 35[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 03/0 : 35[3] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 00/0 : 40[0] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 02/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 00/0 : 38[6] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 01/0 : 34[2] -> 35[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 01/0 : 40[0] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 00/0 : 36[4] -> 37[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 03/0 : 12[4] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 02/0 : 40[0] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 02/0 : 34[2] -> 35[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 00/0 : 42[2] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 00/0 : 32[0] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 01/0 : 38[6] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 01/0 : 36[4] -> 37[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 03/0 : 34[2] -> 35[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 03/0 : 40[0] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 02/0 : 36[4] -> 37[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 02/0 : 38[6] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 01/0 : 42[2] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 01/0 : 32[0] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 00/0 : 41[1] -> 42[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 03/0 : 36[4] -> 37[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 02/0 : 42[2] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 01/0 : 41[1] -> 42[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 02/0 : 32[0] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 03/0 : 38[6] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 03/0 : 42[2] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 02/0 : 41[1] -> 42[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 00/0 : 39[7] -> 40[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 03/0 : 32[0] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 01/0 : 39[7] -> 40[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 00/0 : 43[3] -> 44[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 03/0 : 41[1] -> 42[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 00/0 : 33[1] -> 34[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 00/0 : 34[2] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 01/0 : 43[3] -> 44[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 01/0 : 33[1] -> 34[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 02/0 : 39[7] -> 40[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 02/0 : 43[3] -> 44[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 01/0 : 34[2] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 02/0 : 33[1] -> 34[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 03/0 : 39[7] -> 40[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 03/0 : 43[3] -> 44[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 00/0 : 37[5] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 02/0 : 34[2] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 03/0 : 33[1] -> 34[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 03/0 : 34[2] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 00/0 : 36[4] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 01/0 : 37[5] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 00/0 : 35[3] -> 36[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 02/0 : 37[5] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 01/0 : 36[4] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 01/0 : 35[3] -> 36[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 03/0 : 37[5] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 02/0 : 36[4] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 02/0 : 35[3] -> 36[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 00/0 : 38[6] -> 39[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 03/0 : 35[3] -> 36[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 03/0 : 36[4] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 01/0 : 38[6] -> 39[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 00/0 : 37[5] -> 38[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 02/0 : 38[6] -> 39[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 01/0 : 37[5] -> 38[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 03/0 : 38[6] -> 39[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 02/0 : 37[5] -> 38[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 00/0 : 31[7] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 00/0 : 41[1] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 03/0 : 37[5] -> 38[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 00/0 : 43[3] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 01/0 : 31[7] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 01/0 : 41[1] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 02/0 : 31[7] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 01/0 : 43[3] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 03/0 : 31[7] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 02/0 : 41[1] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 01/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 00/0 : 32[0] -> 33[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 02/0 : 43[3] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 00/0 : 45[5] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 00/0 : 50[2] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 03/0 : 41[1] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 00/0 : 56[0] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 01/0 : 32[0] -> 33[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 03/0 : 43[3] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 00/0 : 42[2] -> 43[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 01/0 : 50[2] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 02/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 02/0 : 32[0] -> 33[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 00/0 : 44[4] -> 45[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 01/0 : 45[5] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 03/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 01/0 : 56[0] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 02/0 : 50[2] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 03/0 : 32[0] -> 33[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 01/0 : 42[2] -> 43[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 01/0 : 44[4] -> 45[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 02/0 : 45[5] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 02/0 : 56[0] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 03/0 : 50[2] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 02/0 : 42[2] -> 43[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 00/0 : 39[7] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 00/0 : 44[4] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 02/0 : 44[4] -> 45[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 00/0 : 60[4] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 03/0 : 45[5] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 00/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 00/0 : 51[3] -> 52[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 01/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 03/0 : 42[2] -> 43[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 03/0 : 56[0] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 00/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 01/0 : 44[4] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 03/0 : 44[4] -> 45[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 00/0 : 46[6] -> 47[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 01/0 : 51[3] -> 52[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 00/0 : 47[7] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 01/0 : 60[4] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 00/0 : 57[1] -> 58[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 01/0 : 39[7] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 02/0 : 44[4] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 00/0 : 49[1] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 01/0 : 46[6] -> 47[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 00/0 : 61[5] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 01/0 : 57[1] -> 58[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 01/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 02/0 : 60[4] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 02/0 : 51[3] -> 52[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 02/0 : 39[7] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 03/0 : 44[4] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 02/0 : 46[6] -> 47[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 02/0 : 57[1] -> 58[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 01/0 : 61[5] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 01/0 : 49[1] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 03/0 : 60[4] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 03/0 : 39[7] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 03/0 : 51[3] -> 52[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 00/0 : 46[6] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 01/0 : 47[7] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 02/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 00/0 : 45[5] -> 46[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 03/0 : 46[6] -> 47[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 03/0 : 57[1] -> 58[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 00/0 : 57[1] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 02/0 : 49[1] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 02/0 : 61[5] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 00/0 : 61[5] -> 62[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 00/0 : 40[0] -> 41[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 01/0 : 45[5] -> 46[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 03/0 : 10[2] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 02/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 02/0 : 47[7] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 01/0 : 46[6] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 03/0 : 49[1] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 03/0 : 61[5] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 01/0 : 40[0] -> 41[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 01/0 : 61[5] -> 62[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 02/0 : 45[5] -> 46[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 03/0 : 47[7] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 00/0 : 50[2] -> 51[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 01/0 : 57[1] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 02/0 : 46[6] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 00/0 : 62[6] -> 63[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 03/0 : 45[5] -> 46[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 02/0 : 40[0] -> 41[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 02/0 : 61[5] -> 62[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 01/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 02/0 : 57[1] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 01/0 : 50[2] -> 51[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 00/0 : 48[0] -> 49[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 03/0 : 46[6] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 01/0 : 62[6] -> 63[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 03/0 : 61[5] -> 62[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 03/0 : 57[1] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 02/0 : 50[2] -> 51[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 02/0 : 62[6] -> 63[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 01/0 : 48[0] -> 49[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 03/0 : 40[0] -> 41[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 00/0 : 47[7] -> 48[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 00/0 : 48[0] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 00/0 : 58[2] -> 59[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 03/0 : 62[6] -> 63[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 01/0 : 47[7] -> 48[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 03/0 : 50[2] -> 51[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 02/0 : 48[0] -> 49[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 00/0 : 31[7] -> 32[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 02/0 : 47[7] -> 48[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 01/0 : 58[2] -> 59[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 03/0 : 48[0] -> 49[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 01/0 : 48[0] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 01/0 : 31[7] -> 32[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 03/0 : 47[7] -> 48[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 02/0 : 48[0] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 02/0 : 58[2] -> 59[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 00/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 03/0 : 48[0] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 02/0 : 31[7] -> 32[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 03/0 : 58[2] -> 59[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 01/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 03/0 : 31[7] -> 32[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 00/0 : 53[5] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 00/0 : 49[1] -> 50[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 00/0 : 58[2] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 01/0 : 49[1] -> 50[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 02/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 00/0 : 59[3] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 01/0 : 53[5] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 01/0 : 58[2] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 02/0 : 49[1] -> 50[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 02/0 : 53[5] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 01/0 : 59[3] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 03/0 : 29[5] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 03/0 : 49[1] -> 50[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 00/0 : 62[6] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 02/0 : 58[2] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 03/0 : 53[5] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 02/0 : 59[3] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 03/0 : 58[2] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 00/0 : 54[6] -> 55[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 02/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 01/0 : 54[6] -> 55[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 03/0 : 59[3] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 01/0 : 62[6] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 00/0 : 59[3] -> 60[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 00/0 : 52[4] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 02/0 : 54[6] -> 55[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 00/0 : 54[6] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 00/0 : 60[4] -> 61[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 03/0 : 54[6] -> 55[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 02/0 : 62[6] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 01/0 : 59[3] -> 60[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 00/0 : 55[7] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 01/0 : 60[4] -> 61[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 01/0 : 52[4] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 01/0 : 54[6] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 03/0 : 62[6] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 02/0 : 59[3] -> 60[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 02/0 : 60[4] -> 61[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 03/0 : 26[2] -> 27[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 00/0 : 51[3] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 02/0 : 52[4] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 02/0 : 54[6] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 01/0 : 55[7] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 00/0 : 63[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 03/0 : 60[4] -> 61[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 03/0 : 59[3] -> 60[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 02/0 : 55[7] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 01/0 : 63[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 03/0 : 54[6] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 03/0 : 52[4] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 03/0 : 55[7] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 02/0 : 63[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 01/0 : 51[3] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 00/0 : 53[5] -> 54[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 00/0 : 56[0] -> 57[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 03/0 : 63[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 00/0 : 55[7] -> 56[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 02/0 : 51[3] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 01/0 : 53[5] -> 54[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 01/0 : 56[0] -> 57[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 01/0 : 55[7] -> 56[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 03/0 : 51[3] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 02/0 : 53[5] -> 54[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 02/0 : 56[0] -> 57[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 03/0 : 56[0] -> 57[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 02/0 : 55[7] -> 56[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 03/0 : 53[5] -> 54[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 00/0 : 52[4] -> 53[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 03/0 : 55[7] -> 56[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 01/0 : 52[4] -> 53[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 02/0 : 52[4] -> 53[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 03/0 : 52[4] -> 53[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 01/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 02/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 03/0 : 7[7] -> 8[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 01/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 00/0 : 63[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 01/0 : 63[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 02/0 : 63[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 03/0 : 63[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 02/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 01/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 02/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 03/0 : 20[4] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 03/0 : 18[2] -> 19[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 01/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 02/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 03/0 : 23[7] -> 24[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Connected all rings +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 00/0 : 44[4] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 01/0 : 44[4] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Connected all rings +ip-26-0-162-46:32942:33312 [1] NCCL INFO Connected all rings +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 02/0 : 41[1] -> 43[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 02/0 : 33[1] -> 35[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 03/0 : 41[1] -> 43[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 03/0 : 33[1] -> 35[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Connected all rings +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Connected all rings +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 02/0 : 39[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 02/0 : 41[1] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 03/0 : 39[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 03/0 : 41[1] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 02/0 : 47[7] -> 55[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 02/0 : 43[3] -> 45[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 03/0 : 47[7] -> 55[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 03/0 : 43[3] -> 45[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 02/0 : 43[3] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 03/0 : 43[3] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Connected all rings +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 00/0 : 42[2] -> 44[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 01/0 : 42[2] -> 44[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Connected all rings +ip-26-0-162-46:32943:33311 [2] NCCL INFO Connected all rings +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 00/0 : 34[2] -> 36[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 01/0 : 34[2] -> 36[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Connected all rings +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Connected all rings +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 00/0 : 36[4] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 02/0 : 43[3] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 01/0 : 36[4] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 00/0 : 36[4] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 03/0 : 43[3] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 01/0 : 36[4] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 00/0 : 40[0] -> 44[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 01/0 : 40[0] -> 44[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 02/0 : 39[7] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 02/0 : 45[5] -> 43[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 03/0 : 39[7] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 03/0 : 45[5] -> 43[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Connected all rings +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 02/0 : 31[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Connected all rings +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Connected all rings +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 03/0 : 31[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 00/0 : 34[2] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 01/0 : 34[2] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 00/0 : 36[4] -> 38[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 01/0 : 36[4] -> 38[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Connected all rings +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 00/0 : 36[4] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 00/0 : 38[6] -> 36[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 00/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 01/0 : 36[4] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 00/0 : 36[4] -> 40[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 01/0 : 10[2] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 01/0 : 38[6] -> 36[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 01/0 : 36[4] -> 40[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 00/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 01/0 : 12[4] -> 14[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 00/0 : 40[0] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 01/0 : 40[0] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Connected all rings +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 00/0 : 16[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 01/0 : 16[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 00/0 : 32[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 01/0 : 32[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Connected all rings +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 00/0 : 50[2] -> 52[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Connected all rings +ip-26-0-162-46:32946:33310 [5] NCCL INFO Connected all rings +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 01/0 : 50[2] -> 52[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 02/0 : 35[3] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 00/0 : 40[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 03/0 : 35[3] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 01/0 : 40[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 00/0 : 48[0] -> 56[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 01/0 : 48[0] -> 56[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Connected all rings +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Connected all rings +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Connected all rings +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 02/0 : 59[3] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 03/0 : 59[3] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Connected all rings +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 00/0 : 42[2] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 01/0 : 42[2] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 00/0 : 44[4] -> 46[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 01/0 : 44[4] -> 46[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 00/0 : 44[4] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 01/0 : 44[4] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 00/0 : 46[6] -> 44[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 00/0 : 40[0] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 01/0 : 46[6] -> 44[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 01/0 : 40[0] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 00/0 : 44[4] -> 40[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 00/0 : 40[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 02/0 : 57[1] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 01/0 : 44[4] -> 40[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 01/0 : 40[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Connected all rings +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 03/0 : 57[1] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 02/0 : 57[1] -> 59[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 02/0 : 59[3] -> 61[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Connected all rings +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 03/0 : 57[1] -> 59[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 03/0 : 59[3] -> 61[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 00/0 : 48[0] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 01/0 : 48[0] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 02/0 : 59[3] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 00/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 03/0 : 59[3] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 01/0 : 28[4] -> 30[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 02/0 : 61[5] -> 59[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 02/0 : 55[7] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 03/0 : 61[5] -> 59[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 03/0 : 55[7] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Connected all rings +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 02/0 : 35[3] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 03/0 : 35[3] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 02/0 : 39[7] -> 43[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 03/0 : 39[7] -> 43[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Connected all rings +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 00/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 01/0 : 26[2] -> 28[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 02/0 : 43[3] -> 39[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 03/0 : 43[3] -> 39[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Connected all rings +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 00/0 : 60[4] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 01/0 : 60[4] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Connected all rings +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 00/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 01/0 : 8[0] -> 12[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Connected all rings +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 00/0 : 58[2] -> 60[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 01/0 : 58[2] -> 60[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Connected all rings +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 02/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Connected all rings +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 03/0 : 9[1] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 02/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 00/0 : 52[4] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 03/0 : 11[3] -> 13[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 01/0 : 52[4] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Connected all rings +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 02/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 03/0 : 25[1] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 02/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 03/0 : 27[3] -> 29[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Connected all rings +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 02/0 : 33[1] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 03/0 : 33[1] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Connected all rings +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 02/0 : 35[3] -> 37[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 03/0 : 35[3] -> 37[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 00/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Connected all rings +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 01/0 : 24[0] -> 28[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 00/0 : 50[2] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 02/0 : 35[3] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 01/0 : 50[2] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 02/0 : 37[5] -> 35[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 03/0 : 35[3] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 02/0 : 35[3] -> 39[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 00/0 : 52[4] -> 54[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 03/0 : 37[5] -> 35[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 03/0 : 35[3] -> 39[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 01/0 : 52[4] -> 54[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 02/0 : 39[7] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 02/0 : 39[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Connected all rings +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 03/0 : 39[7] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 03/0 : 39[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 00/0 : 52[4] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 00/0 : 54[6] -> 52[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 01/0 : 52[4] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 00/0 : 52[4] -> 56[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 02/0 : 51[3] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 01/0 : 54[6] -> 52[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 01/0 : 52[4] -> 56[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 02/0 : 47[7] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 03/0 : 51[3] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 03/0 : 47[7] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Connected all rings +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 00/0 : 58[2] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 01/0 : 58[2] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 00/0 : 60[4] -> 62[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 01/0 : 60[4] -> 62[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 00/0 : 60[4] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 01/0 : 60[4] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 00/0 : 62[6] -> 60[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 00/0 : 56[0] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 01/0 : 62[6] -> 60[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 01/0 : 56[0] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Connected all rings +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 02/0 : 49[1] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 03/0 : 49[1] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 02/0 : 51[3] -> 53[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 03/0 : 51[3] -> 53[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 02/0 : 53[5] -> 51[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Connected all rings +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 03/0 : 53[5] -> 51[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 02/0 : 51[3] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 03/0 : 51[3] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 02/0 : 55[7] -> 59[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 03/0 : 55[7] -> 59[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Connected all rings +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 00/0 : 52[4] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 01/0 : 52[4] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 00/0 : 56[0] -> 60[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 01/0 : 56[0] -> 60[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 02/0 : 59[3] -> 55[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Connected all rings +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 03/0 : 59[3] -> 55[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 02/0 : 31[7] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 00/0 : 56[0] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 03/0 : 31[7] -> 63[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 01/0 : 56[0] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 00/0 : 48[0] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 00/0 : 60[4] -> 56[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 02/0 : 63[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 01/0 : 48[0] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 01/0 : 60[4] -> 56[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 03/0 : 63[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 00/0 : 56[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 01/0 : 56[0] -> 48[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 00/0 : 32[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 01/0 : 32[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 00/0 : 48[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 01/0 : 48[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Connected all rings +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 02/0 : 49[1] -> 51[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 03/0 : 49[1] -> 51[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 02/0 : 51[3] -> 55[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 02/0 : 51[3] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 03/0 : 51[3] -> 55[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 03/0 : 51[3] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 02/0 : 47[7] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 02/0 : 55[7] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 03/0 : 47[7] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 03/0 : 55[7] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 02/0 : 55[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 02/0 : 31[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 03/0 : 55[7] -> 47[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 03/0 : 31[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 02/0 : 47[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 03/0 : 47[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Connected all rings +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Connected all rings +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Connected all rings +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Connected all rings +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 02/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 03/0 : 7[7] -> 11[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Connected all rings +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Connected all rings +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Connected all rings +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Connected all rings +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Connected all rings +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 00/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 00/0 : 32[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 01/0 : 4[4] -> 8[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 01/0 : 32[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 00/0 : 0[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 01/0 : 0[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Connected all rings +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Connected all rings +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Connected all rings +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 02/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 03/0 : 19[3] -> 21[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Connected all rings +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 02/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 03/0 : 17[1] -> 19[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Connected all rings +ip-26-0-161-221:30112:30486 [6] NCCL INFO Connected all rings +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 00/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 01/0 : 18[2] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 00/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 01/0 : 20[4] -> 22[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 00/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 01/0 : 20[4] -> 24[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Connected all rings +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 02/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 03/0 : 19[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 02/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 03/0 : 23[7] -> 27[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 01/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 02/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 00/0 : 16[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 02/0 : 63[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 01/0 : 16[0] -> 32[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 03/0 : 63[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 02/0 : 31[7] -> 63[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 03/0 : 31[7] -> 63[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 00/0 : 32[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 00/0 : 0[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 01/0 : 32[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 01/0 : 0[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 02/0 : 47[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 00/0 : 32[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 03/0 : 47[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 00/0 : 63[7] -> 62[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 01/0 : 32[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 02/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883513 [7] NCCL INFO Channel 01/0 : 63[7] -> 62[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 00/0 : 48[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 01/0 : 48[0] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 00/0 : 32[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 02/0 : 55[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 01/0 : 32[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 03/0 : 55[7] -> 47[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 02/0 : 47[7] -> 39[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 03/0 : 47[7] -> 39[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 00/0 : 56[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 01/0 : 56[0] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 02/0 : 33[1] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 00/0 : 48[0] -> 40[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33307 [0] NCCL INFO Channel 03/0 : 33[1] -> 32[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 01/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 02/0 : 59[3] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 02/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 01/0 : 48[0] -> 40[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 03/0 : 59[3] -> 55[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 00/0 : 47[7] -> 46[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 03/0 : 27[3] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 02/0 : 55[7] -> 51[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 02/0 : 43[3] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Channel 01/0 : 47[7] -> 46[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 02/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 03/0 : 55[7] -> 51[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 02/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 03/0 : 43[3] -> 39[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 03/0 : 23[7] -> 19[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 03/0 : 11[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 02/0 : 39[7] -> 35[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 00/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 03/0 : 39[7] -> 35[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 02/0 : 49[1] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 00/0 : 44[4] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 01/0 : 28[4] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Channel 03/0 : 49[1] -> 48[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 01/0 : 44[4] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 02/0 : 61[5] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 02/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 00/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 02/0 : 53[5] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 00/0 : 55[7] -> 54[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30106:30489 [0] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 00/0 : 40[0] -> 36[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 03/0 : 61[5] -> 59[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 03/0 : 29[5] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 01/0 : 24[0] -> 20[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Channel 01/0 : 55[7] -> 54[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 03/0 : 53[5] -> 51[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 02/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 01/0 : 40[0] -> 36[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 02/0 : 59[3] -> 57[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 02/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 02/0 : 45[5] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 02/0 : 37[5] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 00/0 : 39[7] -> 38[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30487 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 02/0 : 51[3] -> 49[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 03/0 : 21[5] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 03/0 : 59[3] -> 57[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 03/0 : 27[3] -> 25[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 03/0 : 45[5] -> 43[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 03/0 : 37[5] -> 35[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33309 [7] NCCL INFO Channel 01/0 : 39[7] -> 38[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 03/0 : 51[3] -> 49[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 02/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 00/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 02/0 : 43[3] -> 41[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 02/0 : 35[3] -> 33[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 03/0 : 19[3] -> 17[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 00/0 : 62[6] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 01/0 : 30[6] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 00/0 : 46[6] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 03/0 : 35[3] -> 33[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 00/0 : 38[6] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 00/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 00/0 : 59[3] -> 58[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 00/0 : 58[2] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 01/0 : 62[6] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 00/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 02/0 : 41[1] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 03/0 : 43[3] -> 41[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 00/0 : 54[6] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 01/0 : 38[6] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 01/0 : 46[6] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 01/0 : 22[6] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883143:883511 [3] NCCL INFO Channel 01/0 : 59[3] -> 58[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 00/0 : 50[2] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 00/0 : 51[3] -> 50[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 00/0 : 36[4] -> 34[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Channel 03/0 : 41[1] -> 40[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 01/0 : 54[6] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 01/0 : 28[4] -> 26[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 00/0 : 44[4] -> 42[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 00/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 01/0 : 44[4] -> 42[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 01/0 : 58[2] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 02/0 : 62[6] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 00/0 : 38[6] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 01/0 : 50[2] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 01/0 : 36[4] -> 34[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Channel 01/0 : 51[3] -> 50[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 00/0 : 46[6] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 02/0 : 54[6] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 03/0 : 62[6] -> 61[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 02/0 : 58[2] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 00/0 : 43[3] -> 42[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 01/0 : 20[4] -> 18[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 00/0 : 35[3] -> 34[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 02/0 : 50[2] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30490 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 01/0 : 38[6] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 01/0 : 46[6] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 03/0 : 54[6] -> 53[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 02/0 : 61[5] -> 60[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 00/0 : 34[2] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Channel 01/0 : 43[3] -> 42[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 00/0 : 42[2] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 03/0 : 50[2] -> 49[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 03/0 : 58[2] -> 57[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 02/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33314 [3] NCCL INFO Channel 01/0 : 35[3] -> 34[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 00/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 02/0 : 38[6] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 02/0 : 46[6] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 00/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 00/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 01/0 : 34[2] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 00/0 : 47[7] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 02/0 : 53[5] -> 52[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883509 [5] NCCL INFO Channel 03/0 : 61[5] -> 60[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 00/0 : 39[7] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 01/0 : 42[2] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 02/0 : 49[1] -> 48[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 02/0 : 57[1] -> 56[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 03/0 : 38[6] -> 37[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 02/0 : 45[5] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 00/0 : 43[3] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 00/0 : 60[4] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Channel 03/0 : 29[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 02/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 01/0 : 31[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 01/0 : 27[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 02/0 : 34[2] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 03/0 : 46[6] -> 45[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883507 [1] NCCL INFO Channel 03/0 : 57[1] -> 56[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 02/0 : 37[5] -> 36[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 02/0 : 37[5] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 01/0 : 39[7] -> 38[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 01/0 : 47[7] -> 46[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 02/0 : 42[2] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 00/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 01/0 : 23[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 01/0 : 60[4] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 00/0 : 35[3] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Channel 03/0 : 45[5] -> 44[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30111:30485 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 00/0 : 56[0] -> 52[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 03/0 : 34[2] -> 33[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 01/0 : 43[3] -> 42[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 02/0 : 45[5] -> 44[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30110:30488 [4] NCCL INFO Channel 03/0 : 21[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32946:33310 [5] NCCL INFO Channel 03/0 : 37[5] -> 36[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 03/0 : 42[2] -> 41[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 00/0 : 46[6] -> 45[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 01/0 : 56[0] -> 52[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 01/0 : 35[3] -> 34[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32945:33308 [4] NCCL INFO Channel 03/0 : 37[5] -> 36[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 00/0 : 38[6] -> 37[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 02/0 : 33[1] -> 32[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 01/0 : 19[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30107:30492 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 00/0 : 42[2] -> 41[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Channel 03/0 : 45[5] -> 44[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32942:33312 [1] NCCL INFO Channel 03/0 : 33[1] -> 32[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 02/0 : 41[1] -> 40[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 01/0 : 46[6] -> 45[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 01/0 : 38[6] -> 37[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 01/0 : 42[2] -> 41[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 00/0 : 62[6] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 00/0 : 34[2] -> 33[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Channel 03/0 : 41[1] -> 40[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 02/0 : 38[6] -> 37[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 02/0 : 46[6] -> 45[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 01/0 : 34[2] -> 33[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 02/0 : 42[2] -> 41[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 01/0 : 62[6] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33313 [6] NCCL INFO Channel 03/0 : 38[6] -> 37[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 02/0 : 34[2] -> 33[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Channel 03/0 : 46[6] -> 45[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Channel 03/0 : 42[2] -> 41[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30486 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 02/0 : 57[1] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32943:33311 [2] NCCL INFO Channel 03/0 : 34[2] -> 33[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 02/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30108:30491 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 00/0 : 60[4] -> 58[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883506 [0] NCCL INFO Channel 03/0 : 57[1] -> 56[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 03/0 : 13[5] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 01/0 : 60[4] -> 58[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 00/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 01/0 : 12[4] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 00/0 : 63[7] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 00/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Channel 03/0 : 49[1] -> 48[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Channel 03/0 : 53[5] -> 52[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 01/0 : 63[7] -> 62[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 01/0 : 8[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 00/0 : 59[3] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 02/0 : 61[5] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 00/0 : 62[6] -> 61[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883510 [4] NCCL INFO Channel 03/0 : 61[5] -> 60[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 01/0 : 59[3] -> 58[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 01/0 : 62[6] -> 61[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 00/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 02/0 : 62[6] -> 61[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 00/0 : 58[2] -> 57[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 01/0 : 14[6] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883512 [6] NCCL INFO Channel 03/0 : 62[6] -> 61[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 01/0 : 58[2] -> 57[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 02/0 : 58[2] -> 57[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 00/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883142:883508 [2] NCCL INFO Channel 03/0 : 58[2] -> 57[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 01/0 : 12[4] -> 10[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 00/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 01/0 : 15[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 00/0 : 54[6] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 01/0 : 54[6] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 00/0 : 52[4] -> 50[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 01/0 : 52[4] -> 50[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 02/0 : 53[5] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 00/0 : 55[7] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 02/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 00/0 : 51[3] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 03/0 : 11[3] -> 9[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 01/0 : 55[7] -> 54[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Channel 03/0 : 53[5] -> 52[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 01/0 : 51[3] -> 50[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 00/0 : 54[6] -> 53[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 00/0 : 50[2] -> 49[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 01/0 : 54[6] -> 53[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 00/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 01/0 : 50[2] -> 49[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 02/0 : 54[6] -> 53[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 00/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 02/0 : 50[2] -> 49[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 01/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Channel 03/0 : 54[6] -> 53[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 01/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 02/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Channel 03/0 : 50[2] -> 49[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 02/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 03/0 : 14[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 03/0 : 10[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO Connected all trees +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456445:2456813 [0] NCCL INFO comm 0x9146c80 rank 0 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-236:883147:883513 [7] NCCL INFO Connected all trees +ip-26-0-163-236:883147:883513 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883147:883513 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883147:883513 [7] NCCL INFO comm 0x96e3880 rank 63 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO Connected all trees +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32947:33313 [6] NCCL INFO Connected all trees +ip-26-0-162-46:32947:33313 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32947:33313 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO Connected all trees +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO Connected all trees +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO Connected all trees +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461732:2462098 [2] NCCL INFO comm 0x8e52fc0 rank 26 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-162-46:32943:33311 [2] NCCL INFO Connected all trees +ip-26-0-162-46:32943:33311 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32943:33311 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO Connected all trees +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883142:883508 [2] NCCL INFO Connected all trees +ip-26-0-163-236:883142:883508 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883142:883508 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32947:33313 [6] NCCL INFO comm 0x9387380 rank 38 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-178:2582732:2583112 [6] NCCL INFO comm 0x92b4e20 rank 14 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-162-14:2461736:2462095 [6] NCCL INFO comm 0x9eca7c0 rank 30 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO Connected all trees +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461735:2462097 [5] NCCL INFO comm 0x8d3a680 rank 29 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-220:1600809:1601191 [6] NCCL INFO comm 0xa2c8e00 rank 46 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-162-46:32943:33311 [2] NCCL INFO comm 0x8e2a5d0 rank 34 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO Connected all trees +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456447:2456817 [2] NCCL INFO comm 0x8b24900 rank 2 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO Connected all trees +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883141:883507 [1] NCCL INFO Connected all trees +ip-26-0-163-236:883141:883507 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883141:883507 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883142:883508 [2] NCCL INFO comm 0x92b9780 rank 58 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-153:2456446:2456814 [1] NCCL INFO comm 0x8ae4240 rank 1 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO Connected all trees +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600804:1601189 [1] NCCL INFO comm 0xa24efc0 rank 41 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-226:3119069:3119441 [2] NCCL INFO comm 0x8f9ba40 rank 50 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-178:2582727:2583111 [1] NCCL INFO comm 0x8b297f0 rank 9 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-236:883141:883507 [1] NCCL INFO comm 0x913ad70 rank 57 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-226:3119072:3119435 [5] NCCL INFO comm 0xa025180 rank 53 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-236:883146:883512 [6] NCCL INFO Connected all trees +ip-26-0-162-46:32942:33312 [1] NCCL INFO Connected all trees +ip-26-0-162-46:32942:33312 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32942:33312 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883146:883512 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883146:883512 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO Connected all trees +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883146:883512 [6] NCCL INFO comm 0x8789af0 rank 62 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-162-46:32942:33312 [1] NCCL INFO comm 0x87800f0 rank 33 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-153:2456451:2456819 [6] NCCL INFO comm 0xa5138b0 rank 6 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO Connected all trees +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO Connected all trees +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO Connected all trees +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32948:33309 [7] NCCL INFO Connected all trees +ip-26-0-162-46:32948:33309 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32948:33309 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32946:33310 [5] NCCL INFO Connected all trees +ip-26-0-162-46:32945:33308 [4] NCCL INFO Connected all trees +ip-26-0-162-46:32945:33308 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32945:33308 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32946:33310 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32946:33310 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO Connected all trees +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO Connected all trees +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO Connected all trees +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO Connected all trees +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO Connected all trees +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461737:2462096 [7] NCCL INFO comm 0x8a01b30 rank 31 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO Connected all trees +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32944:33314 [3] NCCL INFO Connected all trees +ip-26-0-162-46:32944:33314 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32944:33314 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461734:2462100 [4] NCCL INFO comm 0x9271080 rank 28 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-178:2582733:2583106 [7] NCCL INFO comm 0x92cea80 rank 15 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32948:33309 [7] NCCL INFO comm 0x9e8fff0 rank 39 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO Connected all trees +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO Connected all trees +ip-26-0-162-46:32941:33307 [0] NCCL INFO Connected all trees +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32941:33307 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-162-46:32941:33307 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883143:883511 [3] NCCL INFO Connected all trees +ip-26-0-162-46:32945:33308 [4] NCCL INFO comm 0x950aa70 rank 36 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-236:883143:883511 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883143:883511 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32946:33310 [5] NCCL INFO comm 0x92f53c0 rank 37 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO Connected all trees +ip-26-0-163-236:883145:883509 [5] NCCL INFO Connected all trees +ip-26-0-163-236:883145:883509 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883145:883509 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883144:883510 [4] NCCL INFO Connected all trees +ip-26-0-163-236:883144:883510 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883144:883510 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461731:2462102 [1] NCCL INFO comm 0x87c3ac0 rank 25 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-220:1600806:1601195 [3] NCCL INFO comm 0x9e3f460 rank 43 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO Connected all trees +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456452:2456818 [7] NCCL INFO comm 0x9702ac0 rank 7 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO Connected all trees +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32944:33314 [3] NCCL INFO comm 0xa5fe960 rank 35 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-162-14:2461733:2462101 [3] NCCL INFO comm 0x978e300 rank 27 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-162-14:2461730:2462099 [0] NCCL INFO comm 0x94d23b0 rank 24 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO Connected all trees +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600810:1601193 [7] NCCL INFO comm 0x92efbb0 rank 47 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-236:883140:883506 [0] NCCL INFO Connected all trees +ip-26-0-163-236:883140:883506 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-236:883140:883506 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582729:2583110 [3] NCCL INFO comm 0x87b7ba0 rank 11 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-220:1600808:1601196 [5] NCCL INFO comm 0x88f3960 rank 45 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600807:1601192 [4] NCCL INFO comm 0x9eae400 rank 44 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-220:1600805:1601190 [2] NCCL INFO comm 0xa523e50 rank 42 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-153:2456448:2456816 [3] NCCL INFO comm 0xa36f7b0 rank 3 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-162-46:32941:33307 [0] NCCL INFO comm 0x9447340 rank 32 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-178:2582728:2583105 [2] NCCL INFO comm 0x90a72e0 rank 10 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-236:883143:883511 [3] NCCL INFO comm 0x9c7e370 rank 59 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-236:883145:883509 [5] NCCL INFO comm 0x8a56780 rank 61 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO Connected all trees +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO Connected all trees +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883144:883510 [4] NCCL INFO comm 0x8b57490 rank 60 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-178:2582731:2583109 [5] NCCL INFO comm 0x90fd0c0 rank 13 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO Connected all trees +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456450:2456820 [5] NCCL INFO comm 0x962bdf0 rank 5 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-221:30108:30491 [2] NCCL INFO Connected all trees +ip-26-0-161-221:30108:30491 [2] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30108:30491 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582730:2583108 [4] NCCL INFO comm 0x8bcb770 rank 12 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-153:2456449:2456815 [4] NCCL INFO comm 0x9f96a40 rank 4 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-236:883140:883506 [0] NCCL INFO comm 0x8f73830 rank 56 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO Connected all trees +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO Connected all trees +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO Connected all trees +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600803:1601194 [0] NCCL INFO comm 0x9450990 rank 40 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO Connected all trees +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119070:3119436 [3] NCCL INFO comm 0x8746f20 rank 51 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-178:2582726:2583107 [0] NCCL INFO comm 0x9c2ebc0 rank 8 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-226:3119071:3119437 [4] NCCL INFO comm 0x8dc8a40 rank 52 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-221:30108:30491 [2] NCCL INFO comm 0x996a8e0 rank 18 nranks 64 cudaDev 2 nvmlDev 2 busId 75000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-221:30112:30486 [6] NCCL INFO Connected all trees +ip-26-0-161-221:30112:30486 [6] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30112:30486 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119068:3119440 [1] NCCL INFO comm 0x8f3a9f0 rank 49 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-226:3119067:3119434 [0] NCCL INFO comm 0x99200c0 rank 48 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-226:3119074:3119438 [7] NCCL INFO comm 0xa07bf70 rank 55 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-163-226:3119073:3119439 [6] NCCL INFO comm 0x8a6a4b0 rank 54 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-221:30111:30485 [5] NCCL INFO Connected all trees +ip-26-0-161-221:30111:30485 [5] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30111:30485 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30112:30486 [6] NCCL INFO comm 0x8d03370 rank 22 nranks 64 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-221:30107:30492 [1] NCCL INFO Connected all trees +ip-26-0-161-221:30107:30492 [1] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30107:30492 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30111:30485 [5] NCCL INFO comm 0xa547fc0 rank 21 nranks 64 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-221:30107:30492 [1] NCCL INFO comm 0xa1b8af0 rank 17 nranks 64 cudaDev 1 nvmlDev 1 busId 64000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-221:30109:30490 [3] NCCL INFO Connected all trees +ip-26-0-161-221:30109:30490 [3] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30109:30490 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30109:30490 [3] NCCL INFO comm 0xa39a640 rank 19 nranks 64 cudaDev 3 nvmlDev 3 busId 86000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-221:30113:30487 [7] NCCL INFO Connected all trees +ip-26-0-161-221:30113:30487 [7] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30113:30487 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30110:30488 [4] NCCL INFO Connected all trees +ip-26-0-161-221:30110:30488 [4] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30110:30488 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30106:30489 [0] NCCL INFO Connected all trees +ip-26-0-161-221:30106:30489 [0] NCCL INFO threadThresholds 8/8/64 | 512/8/64 | 512 | 512 +ip-26-0-161-221:30106:30489 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30113:30487 [7] NCCL INFO comm 0x98b0870 rank 23 nranks 64 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-221:30110:30488 [4] NCCL INFO comm 0x898ca30 rank 20 nranks 64 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56edd8b75829b16a - Init COMPLETE +ip-26-0-161-221:30106:30489 [0] NCCL INFO comm 0x8eb32c0 rank 16 nranks 64 cudaDev 0 nvmlDev 0 busId 53000 commId 0x56edd8b75829b16a - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Using network Libfabric +ip-26-0-162-46:32941:33340 [0] NCCL INFO Using network Libfabric +ip-26-0-162-46:32947:33343 [6] NCCL INFO Using network Libfabric +ip-26-0-162-46:32948:33341 [7] NCCL INFO Using network Libfabric +ip-26-0-162-46:32945:33342 [4] NCCL INFO Using network Libfabric +ip-26-0-162-46:32944:33346 [3] NCCL INFO Using network Libfabric +ip-26-0-162-46:32946:33347 [5] NCCL INFO Using network Libfabric +ip-26-0-162-46:32943:33345 [2] NCCL INFO Using network Libfabric +ip-26-0-162-46:32942:33344 [1] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-162-46:32948:33341 [7] NCCL INFO comm 0x9ea5af0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe67f043f3e68418a - Init START +ip-26-0-162-46:32944:33346 [3] NCCL INFO comm 0xa616f40 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe67f043f3e68418a - Init START +ip-26-0-162-46:32945:33342 [4] NCCL INFO comm 0x951fcf0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe67f043f3e68418a - Init START +ip-26-0-162-46:32943:33345 [2] NCCL INFO comm 0x8e445a0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe67f043f3e68418a - Init START +ip-26-0-162-46:32942:33344 [1] NCCL INFO comm 0x8795120 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe67f043f3e68418a - Init START +ip-26-0-162-46:32941:33340 [0] NCCL INFO comm 0x945cbb0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe67f043f3e68418a - Init START +ip-26-0-162-46:32947:33343 [6] NCCL INFO comm 0x939f820 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe67f043f3e68418a - Init START +ip-26-0-162-46:32946:33347 [5] NCCL INFO comm 0x930bf10 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe67f043f3e68418a - Init START +ip-26-0-162-46:32948:33341 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32941:33340 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +NCCL version 2.18.5+cuda12.2 +ip-26-0-162-46:32942:33344 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32945:33342 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32947:33343 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32944:33346 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32943:33345 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32946:33347 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Using network Libfabric +ip-26-0-163-236:883143:883543 [3] NCCL INFO Using network Libfabric +ip-26-0-163-236:883146:883547 [6] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Using network Libfabric +ip-26-0-163-236:883141:883542 [1] NCCL INFO Using network Libfabric +ip-26-0-163-236:883147:883545 [7] NCCL INFO Using network Libfabric +ip-26-0-163-236:883144:883546 [4] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Using network Libfabric +ip-26-0-163-236:883142:883548 [2] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-236:883140:883541 [0] NCCL INFO Using network Libfabric +ip-26-0-163-236:883145:883544 [5] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO comm 0x875c450 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc038bf02e295ae72 - Init START +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO comm 0x8fb1b50 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc038bf02e295ae72 - Init START +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO comm 0x8ddb700 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc038bf02e295ae72 - Init START +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO comm 0xa03bb10 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc038bf02e295ae72 - Init START +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO comm 0x8f4dfa0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc038bf02e295ae72 - Init START +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO comm 0x9932d80 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc038bf02e295ae72 - Init START +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO comm 0x8a7f6e0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc038bf02e295ae72 - Init START +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO comm 0xa0911a0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc038bf02e295ae72 - Init START +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO comm 0x9c42e90 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x92e38bd143182602 - Init START +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO comm 0x92e7620 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x92e38bd143182602 - Init START +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO comm 0x8b3eab0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x92e38bd143182602 - Init START +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO comm 0x8be0790 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x92e38bd143182602 - Init START +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO comm 0x87cd0d0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x92e38bd143182602 - Init START +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO comm 0x92ca150 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x92e38bd143182602 - Init START +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO comm 0x90bc810 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x92e38bd143182602 - Init START +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO comm 0x91112e0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x92e38bd143182602 - Init START +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Using network Libfabric +ip-26-0-163-236:883144:883546 [4] NCCL INFO comm 0x8b71260 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xbc26a6d48baba561 - Init START +ip-26-0-163-236:883140:883541 [0] NCCL INFO comm 0x8f88860 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xbc26a6d48baba561 - Init START +ip-26-0-163-236:883145:883544 [5] NCCL INFO comm 0x8a6cdd0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbc26a6d48baba561 - Init START +ip-26-0-163-236:883143:883543 [3] NCCL INFO comm 0x9c93bb0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbc26a6d48baba561 - Init START +ip-26-0-163-236:883146:883547 [6] NCCL INFO comm 0x879ed20 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbc26a6d48baba561 - Init START +ip-26-0-163-236:883142:883548 [2] NCCL INFO comm 0x92d2de0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xbc26a6d48baba561 - Init START +ip-26-0-163-236:883141:883542 [1] NCCL INFO comm 0x914e530 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbc26a6d48baba561 - Init START +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Using network Libfabric +ip-26-0-163-236:883147:883545 [7] NCCL INFO comm 0x96e6530 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbc26a6d48baba561 - Init START +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Using network Libfabric +ip-26-0-163-236:883144:883546 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883142:883548 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883141:883542 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883147:883545 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO comm 0x8af8460 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x448741a0eba9247c - Init START +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO comm 0x971b360 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x448741a0eba9247c - Init START +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO comm 0x9160ff0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x448741a0eba9247c - Init START +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO comm 0x8b3a260 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x448741a0eba9247c - Init START +ip-26-0-163-236:883146:883547 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883140:883541 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883145:883544 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883143:883543 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO comm 0xa528ae0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x448741a0eba9247c - Init START +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO comm 0xa3849e0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x448741a0eba9247c - Init START +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO comm 0x9641020 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x448741a0eba9247c - Init START +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO comm 0x9fac3e0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x448741a0eba9247c - Init START +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/349 +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO comm 0xa2dd5d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb18e4bd68eb5532b - Init START +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO comm 0x9303de0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb18e4bd68eb5532b - Init START +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO comm 0x8907450 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb18e4bd68eb5532b - Init START +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO comm 0x9ec44a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb18e4bd68eb5532b - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO comm 0xa2638a0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb18e4bd68eb5532b - Init START +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO comm 0x9e53810 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb18e4bd68eb5532b - Init START +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO comm 0x94651e0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb18e4bd68eb5532b - Init START +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO comm 0xa538290 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb18e4bd68eb5532b - Init START +ip-26-0-161-221:30106:30520 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:30111:30523 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:30110:30521 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:30112:30524 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:30108:30527 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:30109:30522 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:30113:30525 [7] NCCL INFO Using network Libfabric +ip-26-0-161-221:30107:30526 [1] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:30106:30520 [0] NCCL INFO comm 0x8ec76c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa1019ab761c69620 - Init START +ip-26-0-161-221:30107:30526 [1] NCCL INFO comm 0xa1ccd20 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa1019ab761c69620 - Init START +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Using network Libfabric +ip-26-0-161-221:30111:30523 [5] NCCL INFO comm 0xa55d9d0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa1019ab761c69620 - Init START +ip-26-0-161-221:30113:30525 [7] NCCL INFO comm 0x98c4c20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa1019ab761c69620 - Init START +ip-26-0-161-221:30112:30524 [6] NCCL INFO comm 0x8d175a0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa1019ab761c69620 - Init START +ip-26-0-161-221:30108:30527 [2] NCCL INFO comm 0x997eb50 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa1019ab761c69620 - Init START +ip-26-0-161-221:30109:30522 [3] NCCL INFO comm 0xa3b0260 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa1019ab761c69620 - Init START +ip-26-0-161-221:30110:30521 [4] NCCL INFO comm 0x89a0c60 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa1019ab761c69620 - Init START +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:30106:30520 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30108:30527 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30107:30526 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30111:30523 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30113:30525 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30112:30524 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30110:30521 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/281 +ip-26-0-161-221:30109:30522 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO comm 0x8d535e0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc4d156db9c0a031b - Init START +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO comm 0x9ee2060 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc4d156db9c0a031b - Init START +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO comm 0x8e6b460 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc4d156db9c0a031b - Init START +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO comm 0x9285d20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc4d156db9c0a031b - Init START +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO comm 0x8a15b60 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc4d156db9c0a031b - Init START +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO comm 0x87d7d60 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc4d156db9c0a031b - Init START +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO comm 0x94e84a0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc4d156db9c0a031b - Init START +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO comm 0x97a36a0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc4d156db9c0a031b - Init START +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-236:883144:883546 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883142:883548 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-236:883143:883543 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-236:883146:883547 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883147:883545 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-162-46:32946:33347 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-236:883141:883542 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-236:883145:883544 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883140:883541 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-236:883141:883542 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-236:883141:883542 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-236:883140:883541 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-236:883140:883541 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883143:883543 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-163-236:883143:883543 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883142:883548 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-236:883142:883548 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883147:883545 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-163-236:883145:883544 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-163-236:883145:883544 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883147:883545 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883146:883547 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-163-236:883146:883547 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883144:883546 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-163-236:883144:883546 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32942:33344 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-162-46:32945:33342 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-162-46:32941:33340 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-162-46:32944:33346 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-46:32947:33343 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-46:32943:33345 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-221:30109:30522 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-162-46:32943:33345 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-162-46:32943:33345 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32942:33344 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-162-46:32942:33344 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32948:33341 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-162-46:32948:33341 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32945:33342 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-162-46:32947:33343 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-162-46:32945:33342 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32944:33346 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-162-46:32947:33343 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32944:33346 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-162-46:32941:33340 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-162-46:32941:33340 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32946:33347 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-162-46:32946:33347 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:30110:30521 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30106:30520 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-221:30107:30526 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-221:30108:30527 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:30112:30524 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30112:30524 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-161-221:30112:30524 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30113:30525 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-161-221:30113:30525 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30107:30526 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:30111:30523 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-161-221:30107:30526 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30109:30522 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-161-221:30110:30521 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-161-221:30111:30523 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30109:30522 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30110:30521 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30108:30527 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-221:30108:30527 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-221:30106:30520 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:30106:30520 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Connected all rings +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Connected all rings +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Connected all rings +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Connected all rings +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Connected all rings +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Connected all rings +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Connected all rings +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Connected all rings +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Connected all rings +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Connected all rings +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Connected all rings +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Connected all rings +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Connected all rings +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Connected all rings +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Connected all rings +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Connected all rings +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Connected all rings +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883147:883545 [7] NCCL INFO Connected all rings +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Connected all rings +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883143:883543 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883140:883541 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883145:883544 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883141:883542 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883546 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883146:883547 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883142:883548 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Connected all rings +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Connected all rings +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Connected all rings +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Trees [0] 1/3/-1->2->4 [1] 1/3/-1->2->4 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 00/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 01/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 02/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 03/04 : 0 1 2 3 4 5 6 7 +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Trees [0] 4/-1/-1->0->-1 [1] 4/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 5/1/-1->3->7 [3] 5/1/-1->3->7 +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 3/-1/-1->7->-1 [3] 3/-1/-1->7->-1 +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Trees [0] 5/7/-1->6->4 [1] 5/7/-1->6->4 [2] -1/-1/-1->6->5 [3] -1/-1/-1->6->5 +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Trees [0] -1/-1/-1->5->6 [1] -1/-1/-1->5->6 [2] 6/4/-1->5->3 [3] 6/4/-1->5->3 +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Trees [0] 2/6/-1->4->0 [1] 2/6/-1->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Connected all rings +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Connected all rings +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Connected all rings +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Connected all rings +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Connected all rings +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Connected all rings +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Connected all rings +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Connected all rings +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Connected all rings +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Connected all rings +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Connected all rings +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Connected all rings +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32941:33340 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Connected all rings +ip-26-0-162-46:32945:33342 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32948:33341 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33346 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33344 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600810:1601230 [7] NCCL INFO comm 0x9303de0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb18e4bd68eb5532b - Init COMPLETE +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO Connected all trees +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461737:2462134 [7] NCCL INFO comm 0x8a15b60 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc4d156db9c0a031b - Init COMPLETE +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Connected all rings +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Connected all rings +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Connected all rings +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Connected all rings +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO Connected all trees +ip-26-0-163-236:883147:883545 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-236:883147:883545 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Connected all rings +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 00/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 01/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 02/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 03/0 : 1[1] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 00/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO Connected all trees +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 01/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Connected all rings +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 02/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 03/0 : 3[3] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 00/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 01/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 02/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 03/0 : 7[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Connected all rings +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Connected all rings +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883147:883545 [7] NCCL INFO comm 0x96e6530 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbc26a6d48baba561 - Init COMPLETE +ip-26-0-162-14:2461730:2462129 [0] NCCL INFO comm 0x94e84a0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc4d156db9c0a031b - Init COMPLETE +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600803:1601226 [0] NCCL INFO comm 0x94651e0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb18e4bd68eb5532b - Init COMPLETE +ip-26-0-163-220:1600804:1601227 [1] NCCL INFO comm 0xa2638a0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb18e4bd68eb5532b - Init COMPLETE +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO Connected all trees +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO Connected all trees +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600809:1601228 [6] NCCL INFO comm 0xa2dd5d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb18e4bd68eb5532b - Init COMPLETE +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO Connected all trees +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO Connected all trees +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600805:1601233 [2] NCCL INFO comm 0xa538290 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb18e4bd68eb5532b - Init COMPLETE +ip-26-0-162-14:2461731:2462136 [1] NCCL INFO comm 0x87d7d60 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc4d156db9c0a031b - Init COMPLETE +ip-26-0-163-236:883140:883541 [0] NCCL INFO Connected all trees +ip-26-0-163-236:883140:883541 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-236:883140:883541 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461736:2462132 [6] NCCL INFO comm 0x9ee2060 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc4d156db9c0a031b - Init COMPLETE +ip-26-0-162-14:2461732:2462130 [2] NCCL INFO comm 0x8e6b460 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc4d156db9c0a031b - Init COMPLETE +ip-26-0-161-178:2582733:2583144 [7] NCCL INFO comm 0x92e7620 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x92e38bd143182602 - Init COMPLETE +ip-26-0-163-236:883140:883541 [0] NCCL INFO comm 0x8f88860 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xbc26a6d48baba561 - Init COMPLETE +ip-26-0-163-236:883141:883542 [1] NCCL INFO Connected all trees +ip-26-0-163-236:883141:883542 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-236:883141:883542 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO Connected all trees +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883146:883547 [6] NCCL INFO Connected all trees +ip-26-0-163-236:883146:883547 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-236:883146:883547 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883142:883548 [2] NCCL INFO Connected all trees +ip-26-0-163-236:883142:883548 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-236:883142:883548 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883141:883542 [1] NCCL INFO comm 0x914e530 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbc26a6d48baba561 - Init COMPLETE +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883146:883547 [6] NCCL INFO comm 0x879ed20 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbc26a6d48baba561 - Init COMPLETE +ip-26-0-161-178:2582726:2583138 [0] NCCL INFO comm 0x9c42e90 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x92e38bd143182602 - Init COMPLETE +ip-26-0-163-236:883142:883548 [2] NCCL INFO comm 0x92d2de0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xbc26a6d48baba561 - Init COMPLETE +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO Connected all trees +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600806:1601232 [3] NCCL INFO comm 0x9e53810 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb18e4bd68eb5532b - Init COMPLETE +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO Connected all trees +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO Connected all trees +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600808:1601229 [5] NCCL INFO comm 0x8907450 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb18e4bd68eb5532b - Init COMPLETE +ip-26-0-162-14:2461735:2462133 [5] NCCL INFO comm 0x8d535e0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc4d156db9c0a031b - Init COMPLETE +ip-26-0-163-220:1600807:1601231 [4] NCCL INFO comm 0x9ec44a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb18e4bd68eb5532b - Init COMPLETE +ip-26-0-162-46:32948:33341 [7] NCCL INFO Connected all trees +ip-26-0-162-46:32948:33341 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-46:32948:33341 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461734:2462131 [4] NCCL INFO comm 0x9285d20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc4d156db9c0a031b - Init COMPLETE +ip-26-0-162-14:2461733:2462135 [3] NCCL INFO comm 0x97a36a0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc4d156db9c0a031b - Init COMPLETE +ip-26-0-162-46:32948:33341 [7] NCCL INFO comm 0x9ea5af0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe67f043f3e68418a - Init COMPLETE +ip-26-0-163-236:883145:883544 [5] NCCL INFO Connected all trees +ip-26-0-163-236:883145:883544 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-236:883145:883544 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO Connected all trees +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO Connected all trees +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO Connected all trees +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883143:883543 [3] NCCL INFO Connected all trees +ip-26-0-163-236:883143:883543 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-236:883143:883543 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32941:33340 [0] NCCL INFO Connected all trees +ip-26-0-162-46:32941:33340 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-46:32941:33340 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883144:883546 [4] NCCL INFO Connected all trees +ip-26-0-163-236:883144:883546 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-236:883144:883546 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582728:2583142 [2] NCCL INFO comm 0x90bc810 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x92e38bd143182602 - Init COMPLETE +ip-26-0-163-236:883145:883544 [5] NCCL INFO comm 0x8a6cdd0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbc26a6d48baba561 - Init COMPLETE +ip-26-0-163-236:883143:883543 [3] NCCL INFO comm 0x9c93bb0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbc26a6d48baba561 - Init COMPLETE +ip-26-0-161-178:2582732:2583140 [6] NCCL INFO comm 0x92ca150 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x92e38bd143182602 - Init COMPLETE +ip-26-0-161-178:2582727:2583143 [1] NCCL INFO comm 0x8b3eab0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x92e38bd143182602 - Init COMPLETE +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO Connected all trees +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30112:30524 [6] NCCL INFO Connected all rings +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO Connected all trees +ip-26-0-162-46:32943:33345 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-46:32943:33345 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32941:33340 [0] NCCL INFO comm 0x945cbb0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe67f043f3e68418a - Init COMPLETE +ip-26-0-162-46:32947:33343 [6] NCCL INFO Connected all trees +ip-26-0-162-46:32947:33343 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-46:32947:33343 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883144:883546 [4] NCCL INFO comm 0x8b71260 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xbc26a6d48baba561 - Init COMPLETE +ip-26-0-162-46:32942:33344 [1] NCCL INFO Connected all trees +ip-26-0-162-46:32942:33344 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-46:32942:33344 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30111:30523 [5] NCCL INFO Connected all rings +ip-26-0-162-46:32946:33347 [5] NCCL INFO Connected all trees +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-46:32946:33347 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Connected all rings +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Connected all rings +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119074:3119469 [7] NCCL INFO comm 0xa0911a0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc038bf02e295ae72 - Init COMPLETE +ip-26-0-161-221:30110:30521 [4] NCCL INFO Connected all rings +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Connected all rings +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32943:33345 [2] NCCL INFO comm 0x8e445a0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe67f043f3e68418a - Init COMPLETE +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32947:33343 [6] NCCL INFO comm 0x939f820 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe67f043f3e68418a - Init COMPLETE +ip-26-0-162-46:32942:33344 [1] NCCL INFO comm 0x8795120 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe67f043f3e68418a - Init COMPLETE +ip-26-0-161-221:30113:30525 [7] NCCL INFO Connected all rings +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30113:30525 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32946:33347 [5] NCCL INFO comm 0x930bf10 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe67f043f3e68418a - Init COMPLETE +ip-26-0-161-221:30109:30522 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30111:30523 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30526 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Connected all rings +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30106:30520 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30110:30521 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30112:30524 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30527 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO Connected all trees +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582729:2583139 [3] NCCL INFO comm 0x87cd0d0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x92e38bd143182602 - Init COMPLETE +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO Connected all trees +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO Connected all trees +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582731:2583145 [5] NCCL INFO comm 0x91112e0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x92e38bd143182602 - Init COMPLETE +ip-26-0-161-178:2582730:2583141 [4] NCCL INFO comm 0x8be0790 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x92e38bd143182602 - Init COMPLETE +ip-26-0-162-46:32944:33346 [3] NCCL INFO Connected all trees +ip-26-0-162-46:32944:33346 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-46:32944:33346 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32945:33342 [4] NCCL INFO Connected all trees +ip-26-0-162-46:32945:33342 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-46:32945:33342 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32944:33346 [3] NCCL INFO comm 0xa616f40 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe67f043f3e68418a - Init COMPLETE +ip-26-0-162-46:32945:33342 [4] NCCL INFO comm 0x951fcf0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe67f043f3e68418a - Init COMPLETE +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO Connected all trees +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119067:3119468 [0] NCCL INFO comm 0x9932d80 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc038bf02e295ae72 - Init COMPLETE +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO Connected all trees +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO Connected all trees +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO Connected all trees +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119068:3119472 [1] NCCL INFO comm 0x8f4dfa0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc038bf02e295ae72 - Init COMPLETE +ip-26-0-163-226:3119073:3119473 [6] NCCL INFO comm 0x8a7f6e0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc038bf02e295ae72 - Init COMPLETE +ip-26-0-163-226:3119069:3119474 [2] NCCL INFO comm 0x8fb1b50 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc038bf02e295ae72 - Init COMPLETE +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Connected all rings +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Connected all rings +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Connected all rings +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Connected all rings +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 00/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 01/0 : 2[2] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 00/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 01/0 : 4[4] -> 6[6] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Connected all rings +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Connected all rings +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 00/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 01/0 : 4[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 00/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 01/0 : 6[6] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 00/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 01/0 : 4[4] -> 2[2] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Connected all rings +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Connected all rings +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 02/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 03/0 : 1[1] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 02/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 03/0 : 3[3] -> 5[5] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 02/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 03/0 : 7[7] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 02/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 03/0 : 3[3] -> 7[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 02/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 03/0 : 5[5] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[1] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 00/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 00/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 01/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 02/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 01/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 03/0 : 6[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 02/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 03/0 : 2[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO Connected all trees +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO Connected all trees +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO Connected all trees +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119072:3119471 [5] NCCL INFO comm 0xa03bb10 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc038bf02e295ae72 - Init COMPLETE +ip-26-0-163-226:3119070:3119475 [3] NCCL INFO comm 0x875c450 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc038bf02e295ae72 - Init COMPLETE +ip-26-0-163-226:3119071:3119470 [4] NCCL INFO comm 0x8ddb700 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc038bf02e295ae72 - Init COMPLETE +ip-26-0-161-221:30113:30525 [7] NCCL INFO Connected all trees +ip-26-0-161-221:30113:30525 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:30113:30525 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30113:30525 [7] NCCL INFO comm 0x98c4c20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa1019ab761c69620 - Init COMPLETE +ip-26-0-161-221:30106:30520 [0] NCCL INFO Connected all trees +ip-26-0-161-221:30106:30520 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:30106:30520 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30112:30524 [6] NCCL INFO Connected all trees +ip-26-0-161-221:30112:30524 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:30112:30524 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30106:30520 [0] NCCL INFO comm 0x8ec76c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa1019ab761c69620 - Init COMPLETE +ip-26-0-161-221:30107:30526 [1] NCCL INFO Connected all trees +ip-26-0-161-221:30107:30526 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:30107:30526 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30108:30527 [2] NCCL INFO Connected all trees +ip-26-0-161-221:30108:30527 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:30108:30527 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO Connected all trees +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30112:30524 [6] NCCL INFO comm 0x8d175a0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa1019ab761c69620 - Init COMPLETE +ip-26-0-161-221:30107:30526 [1] NCCL INFO comm 0xa1ccd20 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa1019ab761c69620 - Init COMPLETE +ip-26-0-161-153:2456452:2456848 [7] NCCL INFO comm 0x971b360 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x448741a0eba9247c - Init COMPLETE +ip-26-0-161-221:30108:30527 [2] NCCL INFO comm 0x997eb50 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa1019ab761c69620 - Init COMPLETE +ip-26-0-161-221:30109:30522 [3] NCCL INFO Connected all trees +ip-26-0-161-221:30109:30522 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:30109:30522 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30111:30523 [5] NCCL INFO Connected all trees +ip-26-0-161-221:30111:30523 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:30111:30523 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30109:30522 [3] NCCL INFO comm 0xa3b0260 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa1019ab761c69620 - Init COMPLETE +ip-26-0-161-221:30110:30521 [4] NCCL INFO Connected all trees +ip-26-0-161-221:30110:30521 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-221:30110:30521 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30111:30523 [5] NCCL INFO comm 0xa55d9d0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa1019ab761c69620 - Init COMPLETE +ip-26-0-161-221:30110:30521 [4] NCCL INFO comm 0x89a0c60 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa1019ab761c69620 - Init COMPLETE +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO Connected all trees +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO Connected all trees +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456445:2456847 [0] NCCL INFO comm 0x9160ff0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x448741a0eba9247c - Init COMPLETE +ip-26-0-161-153:2456446:2456853 [1] NCCL INFO comm 0x8af8460 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x448741a0eba9247c - Init COMPLETE +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO Connected all trees +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO Connected all trees +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456451:2456849 [6] NCCL INFO comm 0xa528ae0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x448741a0eba9247c - Init COMPLETE +ip-26-0-161-153:2456447:2456851 [2] NCCL INFO comm 0x8b3a260 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x448741a0eba9247c - Init COMPLETE +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO Connected all trees +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO Connected all trees +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO Connected all trees +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456450:2456850 [5] NCCL INFO comm 0x9641020 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x448741a0eba9247c - Init COMPLETE +ip-26-0-161-153:2456448:2456852 [3] NCCL INFO comm 0xa3849e0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x448741a0eba9247c - Init COMPLETE +ip-26-0-161-153:2456449:2456854 [4] NCCL INFO comm 0x9fac3e0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x448741a0eba9247c - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Using network Libfabric +ip-26-0-162-46:32945:33379 [4] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Using network Libfabric +ip-26-0-162-46:32943:33380 [2] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:30110:30557 [4] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:30108:30558 [2] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Using network Libfabric +ip-26-0-162-46:32941:33381 [0] NCCL INFO Using network Libfabric +ip-26-0-163-236:883140:883579 [0] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Using network Libfabric +ip-26-0-162-46:32947:33382 [6] NCCL INFO Using network Libfabric +ip-26-0-163-236:883141:883580 [1] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Using network Libfabric +ip-26-0-162-46:32948:33383 [7] NCCL INFO Using network Libfabric +ip-26-0-162-46:32946:33384 [5] NCCL INFO Using network Libfabric +ip-26-0-162-46:32942:33385 [1] NCCL INFO Using network Libfabric +ip-26-0-163-236:883142:883581 [2] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Using network Libfabric +ip-26-0-163-236:883147:883585 [7] NCCL INFO Using network Libfabric +ip-26-0-163-236:883145:883582 [5] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Using network Libfabric +ip-26-0-163-236:883144:883584 [4] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:30113:30561 [7] NCCL INFO Using network Libfabric +ip-26-0-162-46:32944:33386 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Using network Libfabric +ip-26-0-161-221:30109:30563 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:30111:30559 [5] NCCL INFO Using network Libfabric +ip-26-0-161-221:30106:30562 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Using network Libfabric +ip-26-0-163-236:883146:883586 [6] NCCL INFO Using network Libfabric +ip-26-0-163-236:883143:883583 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:30112:30560 [6] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO comm 0xab46100 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x765c8d4eee9f16fd - Init START +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO comm 0xaa70dc0 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x63eae430c828975 - Init START +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO comm 0xbee2710 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x765c8d4eee9f16fd - Init START +ip-26-0-161-221:30108:30558 [2] NCCL INFO comm 0xb6f5b80 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x63eae430c828975 - Init START +ip-26-0-161-221:30110:30557 [4] NCCL INFO comm 0x9cc4740 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x765c8d4eee9f16fd - Init START +ip-26-0-162-46:32945:33379 [4] NCCL INFO comm 0xb289480 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x765c8d4eee9f16fd - Init START +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO comm 0xad1e710 rank 3 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x63eae430c828975 - Init START +ip-26-0-162-46:32943:33380 [2] NCCL INFO comm 0xa1226c0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x63eae430c828975 - Init START +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Using network Libfabric +ip-26-0-161-221:30107:30564 [1] NCCL INFO Using network Libfabric +ip-26-0-162-46:32945:33379 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30110:30557 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/281 +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32943:33380 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30108:30558 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883142:883581 [2] NCCL INFO comm 0xb03c960 rank 3 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x934337a0472b35e - Init START +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO comm 0xb859730 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x934337a0472b35e - Init START +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO comm 0xaff4800 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x934337a0472b35e - Init START +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO comm 0xabdabd0 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x934337a0472b35e - Init START +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883142:883581 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO comm 0xbda8a80 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc6bf208a4a3f0e3c - Init START +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO comm 0xbc44bf0 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4ea3cc19210a6446 - Init START +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO comm 0xaac1d40 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x951adbe01ac3d571 - Init START +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO comm 0xb5bfa60 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4ea3cc19210a6446 - Init START +ip-26-0-163-236:883146:883586 [6] NCCL INFO comm 0xa501c80 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4ea3cc19210a6446 - Init START +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO comm 0xb5788d0 rank 0 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc6bf208a4a3f0e3c - Init START +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO comm 0xb1ffa20 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4ea3cc19210a6446 - Init START +ip-26-0-161-221:30111:30559 [5] NCCL INFO comm 0xb87f890 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc6bf208a4a3f0e3c - Init START +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO comm 0xb047dc0 rank 0 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x951adbe01ac3d571 - Init START +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO comm 0x9beb130 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x951adbe01ac3d571 - Init START +ip-26-0-162-46:32946:33384 [5] NCCL INFO comm 0xa5ed330 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc6bf208a4a3f0e3c - Init START +ip-26-0-163-236:883145:883582 [5] NCCL INFO comm 0xa7d90e0 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x951adbe01ac3d571 - Init START +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO comm 0xaa74cd0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x79e8fc6ec0e7f650 - Init START +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO comm 0xa549c10 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x79e8fc6ec0e7f650 - Init START +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO comm 0xb544580 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x79e8fc6ec0e7f650 - Init START +ip-26-0-163-236:883141:883580 [1] NCCL INFO comm 0xaebd840 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x79e8fc6ec0e7f650 - Init START +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32946:33384 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883146:883586 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883145:883582 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30111:30559 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO comm 0xbe02f80 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc02c216b4436be2c - Init START +ip-26-0-163-236:883141:883580 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO comm 0xabfbaf0 rank 0 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc02c216b4436be2c - Init START +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32948:33383 [7] NCCL INFO comm 0xbc0bd40 rank 2 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc02c216b4436be2c - Init START +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30113:30561 [7] NCCL INFO comm 0xb637280 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc02c216b4436be2c - Init START +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO comm 0xa703700 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x41aab6f45ddf0b6d - Init START +ip-26-0-163-236:883143:883583 [3] NCCL INFO comm 0xb9fefc0 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x41aab6f45ddf0b6d - Init START +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO comm 0xb175300 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x41aab6f45ddf0b6d - Init START +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO comm 0xb5160d0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x41aab6f45ddf0b6d - Init START +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO comm 0xbb7b810 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x579a401203de52d4 - Init START +ip-26-0-162-46:32948:33383 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883140:883579 [0] NCCL INFO comm 0xacf63e0 rank 3 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x579a401203de52d4 - Init START +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO comm 0xb256d10 rank 1 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x579a401203de52d4 - Init START +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO comm 0xb1cf770 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x579a401203de52d4 - Init START +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30113:30561 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO comm 0xab16c00 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x188c16233b4c763f - Init START +ip-26-0-163-236:883144:883584 [4] NCCL INFO comm 0xa8db040 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x188c16233b4c763f - Init START +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO comm 0xafeb080 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x188c16233b4c763f - Init START +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO comm 0xb1e3500 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x188c16233b4c763f - Init START +ip-26-0-163-236:883147:883585 [7] NCCL INFO comm 0xa9d9fd0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x42eb39f66646e32e - Init START +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO comm 0xb21baa0 rank 0 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x42eb39f66646e32e - Init START +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO comm 0xa786640 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x42eb39f66646e32e - Init START +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO comm 0xb072b40 rank 2 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x42eb39f66646e32e - Init START +ip-26-0-163-236:883143:883583 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883140:883579 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883144:883584 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883147:883585 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30109:30563 [3] NCCL INFO comm 0xc11f770 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf9a93c6006b5ebd8 - Init START +ip-26-0-162-46:32944:33386 [3] NCCL INFO comm 0xb8f3960 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf9a93c6006b5ebd8 - Init START +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO comm 0xa4ca110 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf9a93c6006b5ebd8 - Init START +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO comm 0xc2baec0 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf9a93c6006b5ebd8 - Init START +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32944:33386 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30109:30563 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO comm 0xb6a4470 rank 3 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3d249a704ac7b9c0 - Init START +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO comm 0xa638950 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3d249a704ac7b9c0 - Init START +ip-26-0-161-221:30106:30562 [0] NCCL INFO comm 0xa1eac00 rank 1 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3d249a704ac7b9c0 - Init START +ip-26-0-162-46:32941:33381 [0] NCCL INFO comm 0xa73d240 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3d249a704ac7b9c0 - Init START +ip-26-0-162-46:32941:33381 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO comm 0xacba910 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe968733fd63ef026 - Init START +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30106:30562 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO comm 0xaa2ebe0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe968733fd63ef026 - Init START +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/349 +ip-26-0-161-221:30107:30564 [1] NCCL INFO comm 0xbf3f620 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe968733fd63ef026 - Init START +ip-26-0-162-46:32942:33385 [1] NCCL INFO comm 0xa4fd6e0 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe968733fd63ef026 - Init START +ip-26-0-162-46:32942:33385 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30107:30564 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO comm 0xc461360 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7242c0f385e2058b - Init START +ip-26-0-161-221:30112:30560 [6] NCCL INFO comm 0xaa8d140 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7242c0f385e2058b - Init START +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO comm 0xa7ed800 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7242c0f385e2058b - Init START +ip-26-0-162-46:32947:33382 [6] NCCL INFO comm 0xa67bda0 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7242c0f385e2058b - Init START +ip-26-0-162-46:32947:33382 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30112:30560 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32941:33381 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-236:883142:883581 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-46:32944:33386 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883141:883580 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-236:883143:883583 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-236:883144:883584 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-162-46:32945:33379 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-162-46:32948:33383 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:30109:30563 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-221:30109:30563 [3] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:30109:30563 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32944:33386 [3] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-162-46:32944:33386 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32943:33380 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-162-46:32946:33384 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883143:883583 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-236:883143:883583 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32947:33382 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-162-46:32942:33385 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-236:883142:883581 [2] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-236:883142:883581 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-236:883146:883586 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-236:883140:883579 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-236:883140:883579 [0] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-236:883140:883579 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:30108:30558 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:30108:30558 [2] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:30108:30558 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32943:33380 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-162-46:32943:33380 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883145:883582 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883147:883585 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883144:883584 [4] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-236:883144:883584 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883146:883586 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-236:883146:883586 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30113:30561 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30113:30561 [7] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:30113:30561 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32948:33383 [7] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-162-46:32948:33383 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30110:30557 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:30110:30557 [4] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:30110:30557 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32945:33379 [4] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-162-46:32945:33379 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883141:883580 [1] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-236:883141:883580 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30107:30564 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-221:30107:30564 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:30107:30564 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32942:33385 [1] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-162-46:32942:33385 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883145:883582 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-236:883145:883582 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883147:883585 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-236:883147:883585 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30111:30559 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30111:30559 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:30111:30559 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32946:33384 [5] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-162-46:32946:33384 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30112:30560 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30106:30562 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-221:30112:30560 [6] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:30112:30560 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32947:33382 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-162-46:32947:33382 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30106:30562 [0] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-161-221:30106:30562 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32941:33381 [0] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-162-46:32941:33381 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 00/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 01/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 02/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 03/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 00/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 01/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 02/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 03/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 00/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 00/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 01/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 01/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 02/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 02/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 03/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 03/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 00/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 01/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 02/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 03/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 00/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 01/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 02/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 03/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 00/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 01/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 02/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 03/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 00/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 01/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 02/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 03/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 00/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 01/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 02/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 03/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 00/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 01/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 02/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 00/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 03/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 01/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 02/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 03/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 00/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 00/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 01/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 01/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 02/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 02/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 00/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 03/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 01/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 03/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 02/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 03/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 00/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 01/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 00/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 02/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 01/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 02/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 03/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 03/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 00/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 01/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 02/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 03/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 00/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 01/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 02/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 03/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 00/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 01/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 02/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 03/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 00/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 01/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 02/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 03/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 00/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 01/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 02/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 03/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 00/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 01/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 02/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 03/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 00/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 01/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 02/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 03/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 00/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 01/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 02/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 03/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Connected all rings +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Connected all rings +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Connected all rings +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Connected all rings +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33386 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Connected all rings +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 00/0 : 3[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 01/0 : 3[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Connected all rings +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883583 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Connected all rings +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30563 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Connected all rings +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 02/0 : 3[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 03/0 : 3[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Connected all rings +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 02/0 : 3[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 03/0 : 3[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Connected all rings +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 00/0 : 3[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883142:883581 [2] NCCL INFO Channel 01/0 : 3[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Connected all rings +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Connected all rings +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 02/0 : 3[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 03/0 : 3[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 02/0 : 1[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 03/0 : 1[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Connected all rings +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Connected all rings +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Connected all rings +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Connected all rings +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Connected all rings +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Connected all rings +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Connected all rings +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Connected all rings +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 02/0 : 3[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 03/0 : 3[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Connected all rings +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 00/0 : 3[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 01/0 : 3[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33380 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Connected all rings +ip-26-0-162-46:32942:33385 [1] NCCL INFO Connected all rings +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 02/0 : 1[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 03/0 : 1[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 02/0 : 3[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 03/0 : 3[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 00/0 : 3[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883579 [0] NCCL INFO Channel 01/0 : 3[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Connected all rings +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 00/0 : 2[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 01/0 : 2[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 00/0 : 0[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 01/0 : 0[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Connected all rings +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Connected all rings +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Connected all rings +ip-26-0-161-221:30110:30557 [4] NCCL INFO Connected all rings +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Connected all rings +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 00/0 : 3[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 01/0 : 3[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Connected all rings +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 00/0 : 2[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Connected all rings +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 01/0 : 2[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 00/0 : 0[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 01/0 : 0[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Connected all rings +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 00/0 : 0[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33379 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 01/0 : 0[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 00/0 : 2[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 01/0 : 2[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Connected all rings +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Connected all rings +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 00/0 : 2[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 01/0 : 2[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 00/0 : 0[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 01/0 : 0[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Connected all rings +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Connected all rings +ip-26-0-163-236:883146:883586 [6] NCCL INFO Connected all rings +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Connected all rings +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 00/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 01/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 02/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32948:33383 [7] NCCL INFO Channel 03/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 00/0 : 3[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883580 [1] NCCL INFO Channel 01/0 : 3[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883146:883586 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 00/0 : 3[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 00/0 : 3[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 01/0 : 3[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 01/0 : 3[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33385 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Connected all rings +ip-26-0-162-46:32946:33384 [5] NCCL INFO Connected all rings +ip-26-0-162-46:32947:33382 [6] NCCL INFO Connected all rings +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 00/0 : 0[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 01/0 : 0[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 00/0 : 2[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 01/0 : 2[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 00/0 : 0[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 01/0 : 0[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 00/0 : 2[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 01/0 : 2[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 00/0 : 2[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 01/0 : 2[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 00/0 : 0[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 01/0 : 0[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Connected all rings +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 02/0 : 3[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 03/0 : 3[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Connected all rings +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 00/0 : 3[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32946:33384 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 00/0 : 0[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Channel 01/0 : 3[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 01/0 : 0[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 00/0 : 2[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Connected all rings +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 01/0 : 2[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30558 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 00/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 01/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 02/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Channel 03/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Connected all rings +ip-26-0-163-236:883144:883584 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883582 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Connected all rings +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Connected all rings +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 00/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 01/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 02/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 03/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30113:30561 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30557 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Connected all rings +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Connected all rings +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Connected all rings +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 00/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883585 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 01/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 02/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 00/0 : 3[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 03/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Channel 01/0 : 3[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30564 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Connected all rings +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 02/0 : 3[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 03/0 : 3[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 02/0 : 1[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 03/0 : 1[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Connected all rings +ip-26-0-161-221:30111:30559 [5] NCCL INFO Connected all rings +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30559 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Connected all rings +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Connected all rings +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33382 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Connected all rings +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30560 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Connected all rings +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 00/0 : 3[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 01/0 : 3[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33381 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Connected all rings +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 02/0 : 1[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 03/0 : 1[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 02/0 : 3[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 03/0 : 3[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 00/0 : 3[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Channel 01/0 : 3[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30562 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO Connected all trees +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO Connected all trees +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883143:883583 [3] NCCL INFO Connected all trees +ip-26-0-163-236:883143:883583 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-236:883143:883583 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883142:883581 [2] NCCL INFO Connected all trees +ip-26-0-163-236:883142:883581 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-236:883142:883581 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456448:2456898 [3] NCCL INFO comm 0xc2baec0 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf9a93c6006b5ebd8 - Init COMPLETE +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO Connected all trees +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582729:2583184 [3] NCCL INFO comm 0xa703700 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x41aab6f45ddf0b6d - Init COMPLETE +ip-26-0-163-236:883143:883583 [3] NCCL INFO comm 0xb9fefc0 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x41aab6f45ddf0b6d - Init COMPLETE +ip-26-0-163-236:883142:883581 [2] NCCL INFO comm 0xb03c960 rank 3 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x934337a0472b35e - Init COMPLETE +ip-26-0-161-178:2582728:2583185 [2] NCCL INFO comm 0xaff4800 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x934337a0472b35e - Init COMPLETE +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO Connected all trees +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883147:883585 [7] NCCL INFO Connected all trees +ip-26-0-163-236:883147:883585 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-236:883147:883585 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO Connected all trees +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO Connected all trees +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO Connected all trees +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119070:3119512 [3] NCCL INFO comm 0xa4ca110 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf9a93c6006b5ebd8 - Init COMPLETE +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO Connected all trees +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883140:883579 [0] NCCL INFO Connected all trees +ip-26-0-163-236:883140:883579 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-236:883140:883579 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883147:883585 [7] NCCL INFO comm 0xa9d9fd0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x42eb39f66646e32e - Init COMPLETE +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO Connected all trees +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO Connected all trees +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO Connected all trees +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO Connected all trees +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO Connected all trees +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO Connected all trees +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582727:2583187 [1] NCCL INFO comm 0xaa74cd0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x79e8fc6ec0e7f650 - Init COMPLETE +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO Connected all trees +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO Connected all trees +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582732:2583190 [6] NCCL INFO comm 0xb1ffa20 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4ea3cc19210a6446 - Init COMPLETE +ip-26-0-161-178:2582726:2583188 [0] NCCL INFO comm 0xbb7b810 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x579a401203de52d4 - Init COMPLETE +ip-26-0-163-226:3119071:3119506 [4] NCCL INFO comm 0xab46100 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x765c8d4eee9f16fd - Init COMPLETE +ip-26-0-163-236:883140:883579 [0] NCCL INFO comm 0xacf63e0 rank 3 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x579a401203de52d4 - Init COMPLETE +ip-26-0-161-153:2456447:2456886 [2] NCCL INFO comm 0xaa70dc0 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x63eae430c828975 - Init COMPLETE +ip-26-0-162-46:32944:33386 [3] NCCL INFO Connected all trees +ip-26-0-162-46:32944:33386 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-46:32944:33386 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO Connected all trees +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119068:3119511 [1] NCCL INFO comm 0xacba910 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe968733fd63ef026 - Init COMPLETE +ip-26-0-163-236:883145:883582 [5] NCCL INFO Connected all trees +ip-26-0-163-236:883145:883582 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-236:883145:883582 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582731:2583189 [5] NCCL INFO comm 0xb047dc0 rank 0 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x951adbe01ac3d571 - Init COMPLETE +ip-26-0-161-178:2582730:2583182 [4] NCCL INFO comm 0xab16c00 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x188c16233b4c763f - Init COMPLETE +ip-26-0-163-226:3119074:3119510 [7] NCCL INFO comm 0xbe02f80 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc02c216b4436be2c - Init COMPLETE +ip-26-0-161-153:2456452:2456893 [7] NCCL INFO comm 0xabfbaf0 rank 0 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc02c216b4436be2c - Init COMPLETE +ip-26-0-163-226:3119073:3119509 [6] NCCL INFO comm 0xa7ed800 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7242c0f385e2058b - Init COMPLETE +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO Connected all trees +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456450:2456896 [5] NCCL INFO comm 0xb5788d0 rank 0 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc6bf208a4a3f0e3c - Init COMPLETE +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO Connected all trees +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO Connected all trees +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO Connected all trees +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883145:883582 [5] NCCL INFO comm 0xa7d90e0 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x951adbe01ac3d571 - Init COMPLETE +ip-26-0-161-153:2456449:2456888 [4] NCCL INFO comm 0xbee2710 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x765c8d4eee9f16fd - Init COMPLETE +ip-26-0-163-236:883146:883586 [6] NCCL INFO Connected all trees +ip-26-0-163-236:883146:883586 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-236:883146:883586 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119072:3119508 [5] NCCL INFO comm 0xbda8a80 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc6bf208a4a3f0e3c - Init COMPLETE +ip-26-0-162-46:32944:33386 [3] NCCL INFO comm 0xb8f3960 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf9a93c6006b5ebd8 - Init COMPLETE +ip-26-0-161-153:2456445:2456895 [0] NCCL INFO comm 0xa638950 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3d249a704ac7b9c0 - Init COMPLETE +ip-26-0-163-236:883144:883584 [4] NCCL INFO Connected all trees +ip-26-0-163-236:883144:883584 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-236:883144:883584 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119069:3119505 [2] NCCL INFO comm 0xad1e710 rank 3 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x63eae430c828975 - Init COMPLETE +ip-26-0-163-226:3119067:3119507 [0] NCCL INFO comm 0xb6a4470 rank 3 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3d249a704ac7b9c0 - Init COMPLETE +ip-26-0-161-221:30109:30563 [3] NCCL INFO Connected all trees +ip-26-0-161-221:30109:30563 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:30109:30563 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883146:883586 [6] NCCL INFO comm 0xa501c80 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4ea3cc19210a6446 - Init COMPLETE +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO Connected all trees +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883144:883584 [4] NCCL INFO comm 0xa8db040 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x188c16233b4c763f - Init COMPLETE +ip-26-0-161-221:30109:30563 [3] NCCL INFO comm 0xc11f770 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf9a93c6006b5ebd8 - Init COMPLETE +ip-26-0-163-236:883141:883580 [1] NCCL INFO Connected all trees +ip-26-0-163-236:883141:883580 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-236:883141:883580 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456446:2456899 [1] NCCL INFO comm 0xaa2ebe0 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe968733fd63ef026 - Init COMPLETE +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1600805:1601268 [2] NCCL INFO comm 0xb859730 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x934337a0472b35e - Init COMPLETE +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600806:1601272 [3] NCCL INFO comm 0xb175300 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x41aab6f45ddf0b6d - Init COMPLETE +ip-26-0-163-236:883141:883580 [1] NCCL INFO comm 0xaebd840 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x79e8fc6ec0e7f650 - Init COMPLETE +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO Connected all trees +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456451:2456892 [6] NCCL INFO comm 0xc461360 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7242c0f385e2058b - Init COMPLETE +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461732:2462171 [2] NCCL INFO comm 0xabdabd0 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x934337a0472b35e - Init COMPLETE +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO Connected all trees +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600804:1601267 [1] NCCL INFO comm 0xb544580 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x79e8fc6ec0e7f650 - Init COMPLETE +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO Connected all trees +ip-26-0-162-46:32945:33379 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32945:33379 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-46:32945:33379 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO Connected all trees +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32942:33385 [1] NCCL INFO Connected all trees +ip-26-0-162-46:32942:33385 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-46:32942:33385 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600803:1601273 [0] NCCL INFO comm 0xb1cf770 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x579a401203de52d4 - Init COMPLETE +ip-26-0-163-220:1600809:1601270 [6] NCCL INFO comm 0xb5bfa60 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4ea3cc19210a6446 - Init COMPLETE +ip-26-0-162-14:2461733:2462168 [3] NCCL INFO comm 0xb5160d0 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x41aab6f45ddf0b6d - Init COMPLETE +ip-26-0-162-46:32943:33380 [2] NCCL INFO Connected all trees +ip-26-0-162-46:32943:33380 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-46:32943:33380 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32945:33379 [4] NCCL INFO comm 0xb289480 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x765c8d4eee9f16fd - Init COMPLETE +ip-26-0-163-220:1600807:1601269 [4] NCCL INFO comm 0xb1e3500 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x188c16233b4c763f - Init COMPLETE +ip-26-0-162-46:32948:33383 [7] NCCL INFO Connected all trees +ip-26-0-162-46:32948:33383 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-46:32948:33383 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32947:33382 [6] NCCL INFO Connected all trees +ip-26-0-162-46:32947:33382 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-46:32947:33382 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582733:2583192 [7] NCCL INFO comm 0xb21baa0 rank 0 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x42eb39f66646e32e - Init COMPLETE +ip-26-0-162-46:32942:33385 [1] NCCL INFO comm 0xa4fd6e0 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe968733fd63ef026 - Init COMPLETE +ip-26-0-163-220:1600808:1601271 [5] NCCL INFO comm 0x9beb130 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x951adbe01ac3d571 - Init COMPLETE +ip-26-0-162-46:32946:33384 [5] NCCL INFO Connected all trees +ip-26-0-162-46:32946:33384 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-46:32946:33384 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32941:33381 [0] NCCL INFO Connected all trees +ip-26-0-162-46:32941:33381 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-46:32941:33381 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32943:33380 [2] NCCL INFO comm 0xa1226c0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x63eae430c828975 - Init COMPLETE +ip-26-0-161-221:30110:30557 [4] NCCL INFO Connected all trees +ip-26-0-161-221:30110:30557 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:30110:30557 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600810:1601274 [7] NCCL INFO comm 0xb072b40 rank 2 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x42eb39f66646e32e - Init COMPLETE +ip-26-0-162-46:32947:33382 [6] NCCL INFO comm 0xa67bda0 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7242c0f385e2058b - Init COMPLETE +ip-26-0-162-46:32948:33383 [7] NCCL INFO comm 0xbc0bd40 rank 2 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc02c216b4436be2c - Init COMPLETE +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO Connected all trees +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32946:33384 [5] NCCL INFO comm 0xa5ed330 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc6bf208a4a3f0e3c - Init COMPLETE +ip-26-0-161-221:30113:30561 [7] NCCL INFO Connected all trees +ip-26-0-161-221:30113:30561 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:30113:30561 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32941:33381 [0] NCCL INFO comm 0xa73d240 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3d249a704ac7b9c0 - Init COMPLETE +ip-26-0-161-221:30112:30560 [6] NCCL INFO Connected all trees +ip-26-0-161-221:30112:30560 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:30112:30560 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30111:30559 [5] NCCL INFO Connected all trees +ip-26-0-161-221:30111:30559 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:30111:30559 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30107:30564 [1] NCCL INFO Connected all trees +ip-26-0-161-221:30107:30564 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:30107:30564 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO Connected all trees +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30108:30558 [2] NCCL INFO Connected all trees +ip-26-0-161-221:30108:30558 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:30108:30558 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30110:30557 [4] NCCL INFO comm 0x9cc4740 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x765c8d4eee9f16fd - Init COMPLETE +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO Connected all trees +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO Connected all trees +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO Connected all trees +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461730:2462167 [0] NCCL INFO comm 0xb256d10 rank 1 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x579a401203de52d4 - Init COMPLETE +ip-26-0-161-221:30106:30562 [0] NCCL INFO Connected all trees +ip-26-0-161-221:30106:30562 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-221:30106:30562 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30113:30561 [7] NCCL INFO comm 0xb637280 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc02c216b4436be2c - Init COMPLETE +ip-26-0-161-221:30111:30559 [5] NCCL INFO comm 0xb87f890 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc6bf208a4a3f0e3c - Init COMPLETE +ip-26-0-161-221:30112:30560 [6] NCCL INFO comm 0xaa8d140 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7242c0f385e2058b - Init COMPLETE +ip-26-0-161-221:30107:30564 [1] NCCL INFO comm 0xbf3f620 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe968733fd63ef026 - Init COMPLETE +ip-26-0-161-221:30108:30558 [2] NCCL INFO comm 0xb6f5b80 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x63eae430c828975 - Init COMPLETE +ip-26-0-161-221:30106:30562 [0] NCCL INFO comm 0xa1eac00 rank 1 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3d249a704ac7b9c0 - Init COMPLETE +ip-26-0-162-14:2461736:2462173 [6] NCCL INFO comm 0xbc44bf0 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4ea3cc19210a6446 - Init COMPLETE +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO Connected all trees +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461734:2462169 [4] NCCL INFO comm 0xafeb080 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x188c16233b4c763f - Init COMPLETE +ip-26-0-162-14:2461731:2462172 [1] NCCL INFO comm 0xa549c10 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x79e8fc6ec0e7f650 - Init COMPLETE +ip-26-0-162-14:2461735:2462170 [5] NCCL INFO comm 0xaac1d40 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x951adbe01ac3d571 - Init COMPLETE +ip-26-0-162-14:2461737:2462174 [7] NCCL INFO comm 0xa786640 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x42eb39f66646e32e - Init COMPLETE +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO comm 0xa5f6810 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x45350b0012e73b0e - Init START +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO comm 0xab28290 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x45350b0012e73b0e - Init START +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO comm 0xa7cb630 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd8a46ebf12f9bd1d - Init START +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO comm 0xabfefb0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd8a46ebf12f9bd1d - Init START +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO comm 0xb092800 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9f9aebf940842e33 - Init START +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO comm 0xb129200 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9f9aebf940842e33 - Init START +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/349 +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO comm 0xa022860 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb687eca5a67b3139 - Init START +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO comm 0x9fdbef0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb687eca5a67b3139 - Init START +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO comm 0x9caf330 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x123b2a41da96d87c - Init START +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO comm 0xb8697d0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x123b2a41da96d87c - Init START +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO comm 0xa01f480 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd6cab9774e89569b - Init START +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO comm 0xa5a2110 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd6cab9774e89569b - Init START +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO comm 0xb4905a0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa35b1c604b2f67a7 - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO comm 0xa0c2230 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa35b1c604b2f67a7 - Init START +ip-26-0-161-221:30113:30583 [7] NCCL INFO Using network Libfabric +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883147:883604 [7] NCCL INFO Using network Libfabric +ip-26-0-162-46:32948:33406 [7] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Using network Libfabric +ip-26-0-162-46:32942:33411 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO comm 0xa7afdc0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x20e073fed41f0273 - Init START +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO comm 0xba0e5f0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x20e073fed41f0273 - Init START +ip-26-0-162-46:32943:33410 [2] NCCL INFO Using network Libfabric +ip-26-0-162-46:32946:33413 [5] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32947:33415 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Using network Libfabric +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Using network Libfabric +ip-26-0-162-46:32945:33416 [4] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Using network Libfabric +ip-26-0-163-236:883147:883604 [7] NCCL INFO comm 0xb484260 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9caa0c7773c79d28 - Init START +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO comm 0xbe1dd70 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9caa0c7773c79d28 - Init START +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-236:883147:883604 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32944:33418 [3] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO comm 0xb062d70 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd8dbff1639997db0 - Init START +ip-26-0-162-46:32948:33406 [7] NCCL INFO comm 0xbc01080 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd8dbff1639997db0 - Init START +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Using network Libfabric +ip-26-0-162-46:32942:33411 [1] NCCL INFO comm 0xa4f12e0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfe51bdd5ac566c8e - Init START +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO comm 0xbfc05f0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfe51bdd5ac566c8e - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Using network Libfabric +ip-26-0-162-46:32942:33411 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32948:33406 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO comm 0xa664a20 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9094f53be5a6feea - Init START +ip-26-0-161-221:30111:30585 [5] NCCL INFO Using network Libfabric +ip-26-0-162-46:32946:33413 [5] NCCL INFO comm 0xb066d60 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9094f53be5a6feea - Init START +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883145:883605 [5] NCCL INFO Using network Libfabric +ip-26-0-162-46:32943:33410 [2] NCCL INFO comm 0xab9cf20 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf79e35125fbdbdbc - Init START +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO comm 0xc2956b0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf79e35125fbdbdbc - Init START +ip-26-0-163-236:883140:883606 [0] NCCL INFO Using network Libfabric +ip-26-0-162-46:32941:33420 [0] NCCL INFO Using network Libfabric +ip-26-0-162-46:32947:33415 [6] NCCL INFO comm 0xb0f8940 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x85d5756cb3340576 - Init START +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO comm 0xc039000 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x85d5756cb3340576 - Init START +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO comm 0xa7a2fb0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2952cbcd6d678e7e - Init START +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Using network Libfabric +ip-26-0-162-46:32946:33413 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883143:883607 [3] NCCL INFO Using network Libfabric +ip-26-0-162-46:32943:33410 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30113:30583 [7] NCCL INFO comm 0xabe82d0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2952cbcd6d678e7e - Init START +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883141:883608 [1] NCCL INFO Using network Libfabric +ip-26-0-162-46:32947:33415 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO comm 0xbc49170 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x46207d1a61d0d20f - Init START +ip-26-0-162-46:32945:33416 [4] NCCL INFO comm 0xb27db20 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x46207d1a61d0d20f - Init START +ip-26-0-161-221:30113:30583 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32945:33416 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883145:883605 [5] NCCL INFO comm 0xa7f70b0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1199c4f2aa6239b0 - Init START +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO comm 0xbdc4bf0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1199c4f2aa6239b0 - Init START +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Using network Libfabric +ip-26-0-163-236:883143:883607 [3] NCCL INFO comm 0xba1c0d0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe8c4a947d7832db5 - Init START +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO comm 0xa4e7220 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe8c4a947d7832db5 - Init START +ip-26-0-161-221:30107:30589 [1] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Using network Libfabric +ip-26-0-163-236:883141:883608 [1] NCCL INFO comm 0xaed9960 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x78a52f968def7742 - Init START +ip-26-0-163-236:883140:883606 [0] NCCL INFO comm 0xad13770 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb36ae065cba5b2ad - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO comm 0xacd6dd0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x78a52f968def7742 - Init START +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO comm 0xb6bfc20 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb36ae065cba5b2ad - Init START +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO comm 0xb1c4f70 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x998dd2cc01b06b9d - Init START +ip-26-0-163-236:883145:883605 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32941:33420 [0] NCCL INFO comm 0xb1e8990 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x998dd2cc01b06b9d - Init START +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883143:883607 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883141:883608 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30109:30588 [3] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32941:33420 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883140:883606 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO comm 0xaadcfa0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcf7beea71e7aa41f - Init START +ip-26-0-161-221:30111:30585 [5] NCCL INFO comm 0xc2ed030 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcf7beea71e7aa41f - Init START +ip-26-0-161-221:30111:30585 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +NCCL version 2.18.5+cuda12.2 +ip-26-0-161-221:30106:30591 [0] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO comm 0xa5661f0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x351c3a058ff6dbf6 - Init START +ip-26-0-161-221:30107:30589 [1] NCCL INFO comm 0xb4f0ca0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x351c3a058ff6dbf6 - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-236:883142:883609 [2] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30107:30589 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO comm 0xb532db0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4f4f45faeba619be - Init START +ip-26-0-161-221:30109:30588 [3] NCCL INFO comm 0xc116300 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4f4f45faeba619be - Init START +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Using network Libfabric +ip-26-0-161-221:30109:30588 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO comm 0xbbdb170 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf045b207b311fcfd - Init START +ip-26-0-162-46:32944:33418 [3] NCCL INFO comm 0xc370b00 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf045b207b311fcfd - Init START +ip-26-0-162-46:32944:33418 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883146:883610 [6] NCCL INFO Using network Libfabric +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO comm 0xad3b6f0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa304024354a7e588 - Init START +ip-26-0-161-221:30110:30593 [4] NCCL INFO Using network Libfabric +ip-26-0-163-236:883142:883609 [2] NCCL INFO comm 0xb032e70 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa304024354a7e588 - Init START +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883142:883609 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO comm 0xb275940 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9895143590039901 - Init START +ip-26-0-161-221:30106:30591 [0] NCCL INFO comm 0xac596e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9895143590039901 - Init START +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Using network Libfabric +ip-26-0-161-221:30106:30591 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883146:883610 [6] NCCL INFO comm 0xa52a400 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xba8fb14ddbb33d08 - Init START +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO comm 0xa7e1890 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xba8fb14ddbb33d08 - Init START +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883146:883610 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO comm 0xb0139a0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x20d0132a5bbc933 - Init START +ip-26-0-161-221:30110:30593 [4] NCCL INFO comm 0xa70e740 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x20d0132a5bbc933 - Init START +ip-26-0-161-221:30110:30593 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/281 +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Using network Libfabric +ip-26-0-163-236:883144:883611 [4] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-236:883144:883611 [4] NCCL INFO comm 0xa8f7970 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x38d613311471b10c - Init START +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO comm 0xab62470 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x38d613311471b10c - Init START +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-163-236:883144:883611 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30108:30595 [2] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Using network Libfabric +ip-26-0-161-221:30112:30597 [6] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Using network Libfabric +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO comm 0xabf7650 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3d05ef3868edfd23 - Init START +ip-26-0-161-221:30108:30595 [2] NCCL INFO comm 0xb6e5c80 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3d05ef3868edfd23 - Init START +ip-26-0-161-221:30108:30595 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO comm 0xbc6dd80 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8af077eae21a79ff - Init START +ip-26-0-161-221:30112:30597 [6] NCCL INFO comm 0xa03eb70 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8af077eae21a79ff - Init START +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-161-221:30112:30597 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/280 +ip-26-0-162-46:32941:33420 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-162-46:32944:33418 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-46:32942:33411 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-221:30109:30588 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-162-46:32948:33406 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-46:32941:33420 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-162-46:32941:33420 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-162-46:32945:33416 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-46:32945:33416 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-162-46:32945:33416 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:30109:30588 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:30109:30588 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-162-46:32946:33413 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-46:32946:33413 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-162-46:32946:33413 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32943:33410 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-46:32943:33410 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-162-46:32943:33410 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-46:32948:33406 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-162-46:32948:33406 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30108:30595 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-162-46:32947:33415 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-46:32944:33418 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-162-46:32944:33418 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-46:32942:33411 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-162-46:32942:33411 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30110:30593 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-221:30111:30585 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:30111:30585 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:30111:30585 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-46:32947:33415 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-162-46:32947:33415 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:30110:30593 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-221:30110:30593 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-221:30108:30595 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:30108:30595 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-221:30106:30591 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30106:30591 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:30106:30591 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30113:30583 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30113:30583 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:30113:30583 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30112:30597 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30112:30597 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:30112:30597 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30107:30589 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-221:30107:30589 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-221:30107:30589 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32945:33416 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-46:32941:33420 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-221:30109:30588 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32946:33413 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32943:33410 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32948:33406 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32942:33411 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-46:32944:33418 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-221:30111:30585 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-46:32947:33415 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30110:30593 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-221:30108:30595 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30106:30591 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30113:30583 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30112:30597 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-221:30107:30589 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883145:883605 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883145:883605 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-236:883145:883605 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883140:883606 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-236:883140:883606 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-236:883140:883606 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883146:883610 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883146:883610 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-236:883146:883610 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883144:883611 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883144:883611 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-236:883144:883611 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883141:883608 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-236:883141:883608 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-236:883141:883608 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883142:883609 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-236:883142:883609 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-236:883142:883609 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883143:883607 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-236:883143:883607 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-236:883143:883607 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883147:883604 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-236:883147:883604 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-236:883147:883604 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-236:883145:883605 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883605 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883605 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883605 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883605 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883605 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883605 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-236:883145:883605 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883606 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883606 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883606 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883606 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883606 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883606 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883140:883606 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-236:883146:883610 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883140:883606 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883610 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883610 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883610 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883610 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883610 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883610 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-236:883146:883610 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883611 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883611 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883611 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883611 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883611 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883611 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883611 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883144:883611 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-236:883141:883608 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883608 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883608 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883608 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883608 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883608 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883608 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-236:883141:883608 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883609 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883609 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883609 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883609 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883609 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883609 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883609 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-236:883142:883609 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883607 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883607 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883607 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883607 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883607 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883607 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883607 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-236:883143:883607 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30588 [3] NCCL INFO Connected all rings +ip-26-0-161-221:30109:30588 [3] NCCL INFO Connected all trees +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-221:30109:30588 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:30109:30588 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33416 [4] NCCL INFO Connected all rings +ip-26-0-162-46:32945:33416 [4] NCCL INFO Connected all trees +ip-26-0-163-236:883147:883604 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32945:33416 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-46:32945:33416 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883147:883604 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883604 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883604 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883604 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883604 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883604 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-236:883147:883604 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-46:32941:33420 [0] NCCL INFO Connected all rings +ip-26-0-162-46:32941:33420 [0] NCCL INFO Connected all trees +ip-26-0-162-46:32941:33420 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-46:32941:33420 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Connected all rings +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO Connected all trees +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Connected all rings +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32945:33416 [4] NCCL INFO comm 0xb27db20 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x46207d1a61d0d20f - Init COMPLETE +ip-26-0-161-221:30109:30588 [3] NCCL INFO comm 0xc116300 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4f4f45faeba619be - Init COMPLETE +ip-26-0-162-46:32941:33420 [0] NCCL INFO comm 0xb1e8990 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x998dd2cc01b06b9d - Init COMPLETE +ip-26-0-162-14:2461733:2462196 [3] NCCL INFO comm 0xb532db0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4f4f45faeba619be - Init COMPLETE +ip-26-0-163-220:1600803:1601299 [0] NCCL INFO comm 0xb1c4f70 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x998dd2cc01b06b9d - Init COMPLETE +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Connected all rings +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO Connected all trees +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600807:1601297 [4] NCCL INFO comm 0xbc49170 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x46207d1a61d0d20f - Init COMPLETE +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Connected all rings +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO Connected all trees +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582727:2583214 [1] NCCL INFO comm 0xa022860 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb687eca5a67b3139 - Init COMPLETE +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Connected all rings +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO Connected all trees +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32946:33413 [5] NCCL INFO Connected all rings +ip-26-0-162-46:32946:33413 [5] NCCL INFO Connected all trees +ip-26-0-162-46:32946:33413 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-46:32946:33413 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Connected all rings +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO Connected all trees +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Connected all rings +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO Connected all trees +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Connected all rings +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO Connected all trees +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Connected all rings +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32943:33410 [2] NCCL INFO Connected all rings +ip-26-0-162-46:32943:33410 [2] NCCL INFO Connected all trees +ip-26-0-162-46:32943:33410 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-46:32943:33410 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Connected all rings +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Connected all rings +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO Connected all trees +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456446:2456927 [1] NCCL INFO comm 0x9fdbef0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb687eca5a67b3139 - Init COMPLETE +ip-26-0-162-46:32948:33406 [7] NCCL INFO Connected all rings +ip-26-0-162-46:32948:33406 [7] NCCL INFO Connected all trees +ip-26-0-162-46:32948:33406 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-46:32948:33406 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Connected all rings +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO Connected all trees +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Connected all rings +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO Connected all trees +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Connected all rings +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO Connected all trees +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32942:33411 [1] NCCL INFO Connected all rings +ip-26-0-162-46:32942:33411 [1] NCCL INFO Connected all trees +ip-26-0-162-46:32942:33411 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-46:32942:33411 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Connected all rings +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32944:33418 [3] NCCL INFO Connected all rings +ip-26-0-162-46:32944:33418 [3] NCCL INFO Connected all trees +ip-26-0-162-46:32944:33418 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-46:32944:33418 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461735:2462195 [5] NCCL INFO comm 0xaadcfa0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcf7beea71e7aa41f - Init COMPLETE +ip-26-0-162-46:32947:33415 [6] NCCL INFO Connected all rings +ip-26-0-162-46:32947:33415 [6] NCCL INFO Connected all trees +ip-26-0-162-46:32947:33415 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-46:32947:33415 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32946:33413 [5] NCCL INFO comm 0xb066d60 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9094f53be5a6feea - Init COMPLETE +ip-26-0-163-220:1600808:1601295 [5] NCCL INFO comm 0xa664a20 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x9094f53be5a6feea - Init COMPLETE +ip-26-0-161-178:2582733:2583211 [7] NCCL INFO comm 0xa7cb630 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd8a46ebf12f9bd1d - Init COMPLETE +ip-26-0-161-178:2582730:2583215 [4] NCCL INFO comm 0xa0c2230 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa35b1c604b2f67a7 - Init COMPLETE +ip-26-0-162-46:32943:33410 [2] NCCL INFO comm 0xab9cf20 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf79e35125fbdbdbc - Init COMPLETE +ip-26-0-161-153:2456452:2456921 [7] NCCL INFO comm 0xabfefb0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd8a46ebf12f9bd1d - Init COMPLETE +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32948:33406 [7] NCCL INFO comm 0xbc01080 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd8dbff1639997db0 - Init COMPLETE +ip-26-0-163-220:1600804:1601293 [1] NCCL INFO comm 0xbfc05f0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfe51bdd5ac566c8e - Init COMPLETE +ip-26-0-163-220:1600810:1601292 [7] NCCL INFO comm 0xb062d70 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xd8dbff1639997db0 - Init COMPLETE +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Connected all rings +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO Connected all trees +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Connected all rings +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO Connected all trees +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461732:2462200 [2] NCCL INFO comm 0xabf7650 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3d05ef3868edfd23 - Init COMPLETE +ip-26-0-162-46:32942:33411 [1] NCCL INFO comm 0xa4f12e0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfe51bdd5ac566c8e - Init COMPLETE +ip-26-0-161-153:2456445:2456923 [0] NCCL INFO comm 0xb092800 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9f9aebf940842e33 - Init COMPLETE +ip-26-0-162-14:2461734:2462199 [4] NCCL INFO comm 0xb0139a0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x20d0132a5bbc933 - Init COMPLETE +ip-26-0-161-178:2582726:2583212 [0] NCCL INFO comm 0xb129200 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9f9aebf940842e33 - Init COMPLETE +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Connected all rings +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO Connected all trees +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-46:32944:33418 [3] NCCL INFO comm 0xc370b00 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf045b207b311fcfd - Init COMPLETE +ip-26-0-162-46:32947:33415 [6] NCCL INFO comm 0xb0f8940 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x85d5756cb3340576 - Init COMPLETE +ip-26-0-163-220:1600805:1601294 [2] NCCL INFO comm 0xc2956b0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf79e35125fbdbdbc - Init COMPLETE +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Connected all rings +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO Connected all trees +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30111:30585 [5] NCCL INFO Connected all rings +ip-26-0-161-221:30111:30585 [5] NCCL INFO Connected all trees +ip-26-0-161-221:30111:30585 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:30111:30585 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30110:30593 [4] NCCL INFO Connected all rings +ip-26-0-161-221:30110:30593 [4] NCCL INFO Connected all trees +ip-26-0-161-221:30110:30593 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:30110:30593 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30108:30595 [2] NCCL INFO Connected all rings +ip-26-0-161-221:30108:30595 [2] NCCL INFO Connected all trees +ip-26-0-161-221:30108:30595 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:30108:30595 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Connected all rings +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO Connected all trees +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1600806:1601298 [3] NCCL INFO comm 0xbbdb170 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf045b207b311fcfd - Init COMPLETE +ip-26-0-162-14:2461730:2462198 [0] NCCL INFO comm 0xb275940 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9895143590039901 - Init COMPLETE +ip-26-0-163-220:1600809:1601296 [6] NCCL INFO comm 0xc039000 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x85d5756cb3340576 - Init COMPLETE +ip-26-0-161-221:30106:30591 [0] NCCL INFO Connected all rings +ip-26-0-161-221:30106:30591 [0] NCCL INFO Connected all trees +ip-26-0-161-221:30106:30591 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:30106:30591 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461737:2462194 [7] NCCL INFO comm 0xa7a2fb0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2952cbcd6d678e7e - Init COMPLETE +ip-26-0-161-153:2456449:2456931 [4] NCCL INFO comm 0xb4905a0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa35b1c604b2f67a7 - Init COMPLETE +ip-26-0-161-221:30112:30597 [6] NCCL INFO Connected all rings +ip-26-0-161-221:30112:30597 [6] NCCL INFO Connected all trees +ip-26-0-161-221:30112:30597 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:30112:30597 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-162-14:2461736:2462201 [6] NCCL INFO comm 0xbc6dd80 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8af077eae21a79ff - Init COMPLETE +ip-26-0-161-221:30111:30585 [5] NCCL INFO comm 0xc2ed030 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xcf7beea71e7aa41f - Init COMPLETE +ip-26-0-161-221:30110:30593 [4] NCCL INFO comm 0xa70e740 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x20d0132a5bbc933 - Init COMPLETE +ip-26-0-161-221:30108:30595 [2] NCCL INFO comm 0xb6e5c80 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3d05ef3868edfd23 - Init COMPLETE +ip-26-0-162-14:2461731:2462197 [1] NCCL INFO comm 0xa5661f0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x351c3a058ff6dbf6 - Init COMPLETE +ip-26-0-161-221:30106:30591 [0] NCCL INFO comm 0xac596e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9895143590039901 - Init COMPLETE +ip-26-0-161-221:30107:30589 [1] NCCL INFO Connected all rings +ip-26-0-161-221:30107:30589 [1] NCCL INFO Connected all trees +ip-26-0-161-221:30107:30589 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:30107:30589 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30113:30583 [7] NCCL INFO Connected all rings +ip-26-0-161-221:30113:30583 [7] NCCL INFO Connected all trees +ip-26-0-161-221:30113:30583 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-221:30113:30583 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30112:30597 [6] NCCL INFO comm 0xa03eb70 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x8af077eae21a79ff - Init COMPLETE +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Connected all rings +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO Connected all trees +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30107:30589 [1] NCCL INFO comm 0xb4f0ca0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x351c3a058ff6dbf6 - Init COMPLETE +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Connected all rings +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO Connected all trees +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-221:30113:30583 [7] NCCL INFO comm 0xabe82d0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x2952cbcd6d678e7e - Init COMPLETE +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Connected all rings +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO Connected all trees +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582729:2583213 [3] NCCL INFO comm 0x9caf330 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x123b2a41da96d87c - Init COMPLETE +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Connected all rings +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO Connected all trees +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Connected all rings +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO Connected all trees +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582731:2583210 [5] NCCL INFO comm 0xa5f6810 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x45350b0012e73b0e - Init COMPLETE +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Connected all rings +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO Connected all trees +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Connected all rings +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO Connected all trees +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-178:2582732:2583217 [6] NCCL INFO comm 0xa7afdc0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x20e073fed41f0273 - Init COMPLETE +ip-26-0-161-153:2456450:2456920 [5] NCCL INFO comm 0xab28290 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x45350b0012e73b0e - Init COMPLETE +ip-26-0-161-178:2582728:2583216 [2] NCCL INFO comm 0xa5a2110 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd6cab9774e89569b - Init COMPLETE +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Connected all rings +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO Connected all trees +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-153:2456448:2456925 [3] NCCL INFO comm 0xb8697d0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x123b2a41da96d87c - Init COMPLETE +ip-26-0-161-153:2456447:2456932 [2] NCCL INFO comm 0xa01f480 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd6cab9774e89569b - Init COMPLETE +ip-26-0-161-153:2456451:2456934 [6] NCCL INFO comm 0xba0e5f0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x20e073fed41f0273 - Init COMPLETE +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Connected all rings +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO Connected all trees +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883145:883605 [5] NCCL INFO Connected all rings +ip-26-0-163-236:883145:883605 [5] NCCL INFO Connected all trees +ip-26-0-163-236:883145:883605 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-236:883145:883605 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119072:3119536 [5] NCCL INFO comm 0xbdc4bf0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1199c4f2aa6239b0 - Init COMPLETE +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Connected all rings +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO Connected all trees +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883145:883605 [5] NCCL INFO comm 0xa7f70b0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1199c4f2aa6239b0 - Init COMPLETE +ip-26-0-163-236:883140:883606 [0] NCCL INFO Connected all rings +ip-26-0-163-236:883140:883606 [0] NCCL INFO Connected all trees +ip-26-0-163-236:883140:883606 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-236:883140:883606 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Connected all rings +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO Connected all trees +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119067:3119538 [0] NCCL INFO comm 0xb6bfc20 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb36ae065cba5b2ad - Init COMPLETE +ip-26-0-163-236:883140:883606 [0] NCCL INFO comm 0xad13770 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb36ae065cba5b2ad - Init COMPLETE +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Connected all rings +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO Connected all trees +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883146:883610 [6] NCCL INFO Connected all rings +ip-26-0-163-236:883146:883610 [6] NCCL INFO Connected all trees +ip-26-0-163-236:883146:883610 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-236:883146:883610 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Connected all rings +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO Connected all trees +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883141:883608 [1] NCCL INFO Connected all rings +ip-26-0-163-236:883141:883608 [1] NCCL INFO Connected all trees +ip-26-0-163-236:883141:883608 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-236:883141:883608 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883144:883611 [4] NCCL INFO Connected all rings +ip-26-0-163-236:883144:883611 [4] NCCL INFO Connected all trees +ip-26-0-163-236:883144:883611 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-236:883144:883611 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119073:3119545 [6] NCCL INFO comm 0xa7e1890 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xba8fb14ddbb33d08 - Init COMPLETE +ip-26-0-163-226:3119071:3119547 [4] NCCL INFO comm 0xab62470 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x38d613311471b10c - Init COMPLETE +ip-26-0-163-226:3119068:3119541 [1] NCCL INFO comm 0xacd6dd0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x78a52f968def7742 - Init COMPLETE +ip-26-0-163-236:883146:883610 [6] NCCL INFO comm 0xa52a400 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xba8fb14ddbb33d08 - Init COMPLETE +ip-26-0-163-236:883141:883608 [1] NCCL INFO comm 0xaed9960 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x78a52f968def7742 - Init COMPLETE +ip-26-0-163-236:883144:883611 [4] NCCL INFO comm 0xa8f7970 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x38d613311471b10c - Init COMPLETE +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Connected all rings +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO Connected all trees +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883142:883609 [2] NCCL INFO Connected all rings +ip-26-0-163-236:883142:883609 [2] NCCL INFO Connected all trees +ip-26-0-163-236:883142:883609 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-236:883142:883609 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Connected all rings +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO Connected all trees +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119069:3119543 [2] NCCL INFO comm 0xad3b6f0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa304024354a7e588 - Init COMPLETE +ip-26-0-163-236:883143:883607 [3] NCCL INFO Connected all rings +ip-26-0-163-236:883143:883607 [3] NCCL INFO Connected all trees +ip-26-0-163-236:883143:883607 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-236:883143:883607 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Connected all rings +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO Connected all trees +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883147:883604 [7] NCCL INFO Connected all rings +ip-26-0-163-236:883147:883604 [7] NCCL INFO Connected all trees +ip-26-0-163-236:883147:883604 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-236:883147:883604 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-236:883142:883609 [2] NCCL INFO comm 0xb032e70 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa304024354a7e588 - Init COMPLETE +ip-26-0-163-226:3119074:3119533 [7] NCCL INFO comm 0xbe1dd70 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9caa0c7773c79d28 - Init COMPLETE +ip-26-0-163-236:883143:883607 [3] NCCL INFO comm 0xba1c0d0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe8c4a947d7832db5 - Init COMPLETE +ip-26-0-163-226:3119070:3119539 [3] NCCL INFO comm 0xa4e7220 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe8c4a947d7832db5 - Init COMPLETE +ip-26-0-163-236:883147:883604 [7] NCCL INFO comm 0xb484260 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9caa0c7773c79d28 - Init COMPLETE +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + +trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator +self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ +optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( +optimizer = ZeroDistributedOptimizer( File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + +optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( +optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +optimizer = optimizer_builder(param_groups_in_rank) +trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + result = OptimizerFromGradientAccumulator( + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + + result = OptimizerFromGradientAccumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = ZeroDistributedOptimizer( + optimizer = ZeroDistributedOptimizer( File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( +gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(param_groups_in_rank) + optimizer = optimizer_builder(param_groups_in_rank) File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = ZeroDistributedOptimizer( +contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 53.29 GiB. GPU 0 has a total capacty of 79.33 GiB of which 50.84 GiB is free. Including non-PyTorch memory, this process has 28.48 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( +gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda")gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 53.29 GiB. GPU 7 has a total capacty of 79.33 GiB of which 50.84 GiB is free. Including non-PyTorch memory, this process has 28.48 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( +torch.cuda. File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +OutOfMemoryError: CUDA out of memory. Tried to allocate 53.29 GiB. GPU 6 has a total capacty of 79.33 GiB of which 50.88 GiB is free. Including non-PyTorch memory, this process has 28.44 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda")contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + +torch.cuda.torch.cudaOutOfMemoryError.: contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda")CUDA out of memory. Tried to allocate 53.29 GiB. GPU 7 has a total capacty of 79.33 GiB of which 50.84 GiB is free. Including non-PyTorch memory, this process has 28.48 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFOutOfMemoryError + +: CUDA out of memory. Tried to allocate 53.29 GiB. GPU 4 has a total capacty of 79.33 GiB of which 50.76 GiB is free. Including non-PyTorch memory, this process has 28.56 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFcontiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + +contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cudatorch.cuda..torch.cuda OutOfMemoryError.contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda")OutOfMemoryErrorcontiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda"): +: +CUDA out of memory. Tried to allocate 53.29 GiB. GPU 1 has a total capacty of 79.33 GiB of which 50.88 GiB is free. Including non-PyTorch memory, this process has 28.44 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFOutOfMemoryErrorCUDA out of memory. Tried to allocate 53.29 GiB. GPU 0 has a total capacty of 79.33 GiB of which 50.84 GiB is free. Including non-PyTorch memory, this process has 28.48 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +: torch.cudaCUDA out of memory. Tried to allocate 53.29 GiB. GPU 3 has a total capacty of 79.33 GiB of which 50.76 GiB is free. Including non-PyTorch memory, this process has 28.56 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +.torch.cudaOutOfMemoryError.: OutOfMemoryErrorCUDA out of memory. Tried to allocate 53.29 GiB. GPU 6 has a total capacty of 79.33 GiB of which 50.88 GiB is free. Including non-PyTorch memory, this process has 28.44 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF: +CUDA out of memory. Tried to allocate 53.29 GiB. GPU 2 has a total capacty of 79.33 GiB of which 50.88 GiB is free. Including non-PyTorch memory, this process has 28.44 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 53.29 GiB. GPU 4 has a total capacty of 79.33 GiB of which 50.76 GiB is free. Including non-PyTorch memory, this process has 28.56 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 53.29 GiB. GPU 2 has a total capacty of 79.33 GiB of which 50.88 GiB is free. Including non-PyTorch memory, this process has 28.44 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 53.29 GiB. GPU 3 has a total capacty of 79.33 GiB of which 50.76 GiB is free. Including non-PyTorch memory, this process has 28.56 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 53.29 GiB. GPU 5 has a total capacty of 79.33 GiB of which 50.88 GiB is free. Including non-PyTorch memory, this process has 28.44 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 53.29 GiB. GPU 1 has a total capacty of 79.33 GiB of which 50.88 GiB is free. Including non-PyTorch memory, this process has 28.44 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers(optimizer = optimizer_builder(param_groups_in_rank) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 50.03 GiB is free. Including non-PyTorch memory, this process has 29.29 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 145.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + result = OptimizerFromGradientAccumulator(gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +torch.cuda.OutOfMemoryError File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 49.99 GiB is free. Including non-PyTorch memory, this process has 29.32 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 145.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 49.99 GiB is free. Including non-PyTorch memory, this process has 29.32 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 145.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 53.29 GiB. GPU 5 has a total capacty of 79.33 GiB of which 50.88 GiB is free. Including non-PyTorch memory, this process has 28.44 GiB memory in use. Of the allocated memory 26.64 GiB is allocated by PyTorch, and 160.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 49.92 GiB is free. Including non-PyTorch memory, this process has 29.40 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 145.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ +optimizer = ZeroDistributedOptimizer( +gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( +self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) +contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 50.03 GiB is free. Including non-PyTorch memory, this process has 29.29 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 145.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 49.92 GiB is free. Including non-PyTorch memory, this process has 29.40 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 145.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFgradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + +gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 50.03 GiB is free. Including non-PyTorch memory, this process has 29.29 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 145.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 50.03 GiB is free. Including non-PyTorch memory, this process has 29.29 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 145.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = ZeroDistributedOptimizer(optimizer = ZeroDistributedOptimizer( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = ZeroDistributedOptimizer(result = OptimizerFromGradientAccumulator( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +gradient_accumulator = gradient_accumulator_builder(name_to_param.items())gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in +gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 49.98 GiB is free. Including non-PyTorch memory, this process has 29.34 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 161.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator(gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + +torch.cuda. File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 50.02 GiB is free. Including non-PyTorch memory, this process has 29.30 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 161.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) +trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) +trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(param_groups_in_rank) + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda")self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers(contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.torch.cudaOutOfMemoryError.: OutOfMemoryErrorCUDA out of memory. Tried to allocate 55.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 49.90 GiB is free. Including non-PyTorch memory, this process has 29.42 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 161.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF: +CUDA out of memory. Tried to allocate 55.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 49.90 GiB is free. Including non-PyTorch memory, this process has 29.42 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 161.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda. self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers(OutOfMemoryError +: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 50.02 GiB is free. Including non-PyTorch memory, this process has 29.30 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 161.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 50.02 GiB is free. Including non-PyTorch memory, this process has 29.30 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 161.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 50.02 GiB is free. Including non-PyTorch memory, this process has 29.30 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 161.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 49.98 GiB is free. Including non-PyTorch memory, this process has 29.34 GiB memory in use. Of the allocated memory 27.50 GiB is allocated by PyTorch, and 161.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator(optimizer = optimizer_builder(param_groups_in_rank) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(param_groups_in_rank) + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + result = OptimizerFromGradientAccumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +result = OptimizerFromGradientAccumulator( + + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator(result = OptimizerFromGradientAccumulator( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator( + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 49.53 GiB is free. Including non-PyTorch memory, this process has 29.79 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 49.49 GiB is free. Including non-PyTorch memory, this process has 29.83 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 49.41 GiB is free. Including non-PyTorch memory, this process has 29.91 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 1 has a total capacty of 79.33 GiB of which 49.53 GiB is free. Including non-PyTorch memory, this process has 29.79 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 49.53 GiB is free. Including non-PyTorch memory, this process has 29.79 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 49.53 GiB is free. Including non-PyTorch memory, this process has 29.79 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 49.49 GiB is free. Including non-PyTorch memory, this process has 29.83 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 49.53 GiB is free. Including non-PyTorch memory, this process has 29.79 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 49.41 GiB is free. Including non-PyTorch memory, this process has 29.91 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 5 has a total capacty of 79.33 GiB of which 49.53 GiB is free. Including non-PyTorch memory, this process has 29.79 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 0 has a total capacty of 79.33 GiB of which 49.64 GiB is free. Including non-PyTorch memory, this process has 29.68 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cudatorch.cuda. .self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers(OutOfMemoryErrorOutOfMemoryError: : +CUDA out of memory. Tried to allocate 56.00 GiB. GPU 3 has a total capacty of 79.33 GiB of which 49.41 GiB is free. Including non-PyTorch memory, this process has 29.91 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFCUDA out of memory. Tried to allocate 56.00 GiB. GPU 6 has a total capacty of 79.33 GiB of which 49.53 GiB is free. Including non-PyTorch memory, this process has 29.79 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 7 has a total capacty of 79.33 GiB of which 49.49 GiB is free. Including non-PyTorch memory, this process has 29.83 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 4 has a total capacty of 79.33 GiB of which 49.41 GiB is free. Including non-PyTorch memory, this process has 29.91 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 56.00 GiB. GPU 2 has a total capacty of 79.33 GiB of which 49.53 GiB is free. Including non-PyTorch memory, this process has 29.79 GiB memory in use. Of the allocated memory 28.00 GiB is allocated by PyTorch, and 193.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + trainer = DistributedTrainer(config_file) File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) +trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator +self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + trainer = DistributedTrainer(config_file) +result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( +contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 0 has a total capacty of 79.33 GiB of which 50.38 GiB is free. Including non-PyTorch memory, this process has 28.94 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 4 has a total capacty of 79.33 GiB of which 50.30 GiB is free. Including non-PyTorch memory, this process has 29.02 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 3 has a total capacty of 79.33 GiB of which 50.30 GiB is free. Including non-PyTorch memory, this process has 29.02 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + trainer = DistributedTrainer(config_file)trainer = DistributedTrainer(config_file)trainer = DistributedTrainer(config_file) + + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ +trainer = DistributedTrainer(config_file) trainer = DistributedTrainer(config_file) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator +self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer(optimizer = ZeroDistributedOptimizer( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = ZeroDistributedOptimizer(optimizer = ZeroDistributedOptimizer( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +optimizer = optimizer_builder(param_groups_in_rank) +optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder +optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + result = OptimizerFromGradientAccumulator(result = OptimizerFromGradientAccumulator( + +result = OptimizerFromGradientAccumulator( File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) +self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda")result = OptimizerFromGradientAccumulator( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 1 has a total capacty of 79.33 GiB of which 50.42 GiB is free. Including non-PyTorch memory, this process has 28.90 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError : gradient_accumulator = gradient_accumulator_builder(name_to_param.items())CUDA out of memory. Tried to allocate 54.29 GiB. GPU 5 has a total capacty of 79.33 GiB of which 50.42 GiB is free. Including non-PyTorch memory, this process has 28.90 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ +gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( +gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 184, in __init__ + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + self.optimizer, self.grad_accumulator = init_optimizer_and_grad_accumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 383, in init_optimizer_and_grad_accumulator + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + optimizer = ZeroDistributedOptimizer( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/zero.py", line 91, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + optimizer = optimizer_builder(param_groups_in_rank) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 364, in grad_optimizer_builder + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 2 has a total capacty of 79.33 GiB of which 50.42 GiB is free. Including non-PyTorch memory, this process has 28.90 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 6 has a total capacty of 79.33 GiB of which 50.42 GiB is free. Including non-PyTorch memory, this process has 28.90 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 7 has a total capacty of 79.33 GiB of which 50.38 GiB is free. Including non-PyTorch memory, this process has 28.94 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda")torch.cuda +.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 2 has a total capacty of 79.33 GiB of which 50.42 GiB is free. Including non-PyTorch memory, this process has 28.90 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 6 has a total capacty of 79.33 GiB of which 50.42 GiB is free. Including non-PyTorch memory, this process has 28.90 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 7 has a total capacty of 79.33 GiB of which 50.53 GiB is free. Including non-PyTorch memory, this process has 28.79 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda")torch.cuda. +OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 3 has a total capacty of 79.33 GiB of which 50.30 GiB is free. Including non-PyTorch memory, this process has 29.02 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 4 has a total capacty of 79.33 GiB of which 50.30 GiB is free. Including non-PyTorch memory, this process has 29.02 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 0 has a total capacty of 79.33 GiB of which 50.38 GiB is free. Including non-PyTorch memory, this process has 28.94 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 1 has a total capacty of 79.33 GiB of which 50.42 GiB is free. Including non-PyTorch memory, this process has 28.90 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + result = OptimizerFromGradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/optimizer_from_gradient_accumulator.py", line 32, in __init__ + gradient_accumulator = gradient_accumulator_builder(name_to_param.items()) + File "/fsx/nouamane/projects/nanotron/src/nanotron/helpers.py", line 365, in + gradient_accumulator_builder=lambda named_params: FP32GradientAccumulator( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 78, in __init__ + self.fp32_grad_buffers, self._contiguous_fp32_grad_buffer = self.build_grad_buffers( + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 173, in build_grad_buffers + contiguous_buffer_f32_gradients = torch.zeros(needed_buffer_size, dtype=torch.float, device="cuda") +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 54.29 GiB. GPU 5 has a total capacty of 79.33 GiB of which 50.42 GiB is free. Including non-PyTorch memory, this process has 28.90 GiB memory in use. Of the allocated memory 27.14 GiB is allocated by PyTorch, and 161.99 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[2024-12-27 18:42:52,210] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2456445) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 18:42:52,213] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 32941) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 18:42:52,212] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1600806 closing signal SIGTERM +[2024-12-27 18:42:52,212] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1600807 closing signal SIGTERM +[2024-12-27 18:42:52,212] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1600809 closing signal SIGTERM +[2024-12-27 18:42:52,212] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1600810 closing signal SIGTERM +[2024-12-27 18:42:52,214] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2582726) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 18:42:52,219] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 883140) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 18:42:52,220] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 30106) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 18:42:52,221] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 3119067) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-27 18:42:52,222] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2461730) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-178.ec2.internal + rank : 9 (local_rank: 1) + exitcode : 1 (pid: 2582727) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-178.ec2.internal + rank : 10 (local_rank: 2) + exitcode : 1 (pid: 2582728) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-178.ec2.internal + rank : 11 (local_rank: 3) + exitcode : 1 (pid: 2582729) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-178.ec2.internal + rank : 12 (local_rank: 4) + exitcode : 1 (pid: 2582730) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-178.ec2.internal + rank : 13 (local_rank: 5) + exitcode : 1 (pid: 2582731) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-178.ec2.internal + rank : 14 (local_rank: 6) + exitcode : 1 (pid: 2582732) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-178.ec2.internal + rank : 15 (local_rank: 7) + exitcode : 1 (pid: 2582733) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-178.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 2582726) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-46.ec2.internal + rank : 33 (local_rank: 1) + exitcode : 1 (pid: 32942) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-46.ec2.internal + rank : 34 (local_rank: 2) + exitcode : 1 (pid: 32943) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-46.ec2.internal + rank : 35 (local_rank: 3) + exitcode : 1 (pid: 32944) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-46.ec2.internal + rank : 36 (local_rank: 4) + exitcode : 1 (pid: 32945) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-46.ec2.internal + rank : 37 (local_rank: 5) + exitcode : 1 (pid: 32946) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-46.ec2.internal + rank : 38 (local_rank: 6) + exitcode : 1 (pid: 32947) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-46.ec2.internal + rank : 39 (local_rank: 7) + exitcode : 1 (pid: 32948) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-46.ec2.internal + rank : 32 (local_rank: 0) + exitcode : 1 (pid: 32941) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-221.ec2.internal + rank : 17 (local_rank: 1) + exitcode : 1 (pid: 30107) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-221.ec2.internal + rank : 18 (local_rank: 2) + exitcode : 1 (pid: 30108) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-221.ec2.internal + rank : 19 (local_rank: 3) + exitcode : 1 (pid: 30109) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-221.ec2.internal + rank : 20 (local_rank: 4) + exitcode : 1 (pid: 30110) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-221.ec2.internal + rank : 21 (local_rank: 5) + exitcode : 1 (pid: 30111) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-221.ec2.internal + rank : 22 (local_rank: 6) + exitcode : 1 (pid: 30112) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-221.ec2.internal + rank : 23 (local_rank: 7) + exitcode : 1 (pid: 30113) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-221.ec2.internal + rank : 16 (local_rank: 0) + exitcode : 1 (pid: 30106) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-153.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 2456446) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-153.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 2456447) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-153.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 2456448) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-153.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 2456449) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-153.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 2456450) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-153.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 2456451) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-153.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 2456452) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_18:42:52 + host : ip-26-0-161-153.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 2456445) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-236.ec2.internal + rank : 57 (local_rank: 1) + exitcode : 1 (pid: 883141) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-236.ec2.internal + rank : 58 (local_rank: 2) + exitcode : 1 (pid: 883142) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-236.ec2.internal + rank : 59 (local_rank: 3) + exitcode : 1 (pid: 883143) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-236.ec2.internal + rank : 60 (local_rank: 4) + exitcode : 1 (pid: 883144) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-236.ec2.internal + rank : 61 (local_rank: 5) + exitcode : 1 (pid: 883145) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-236.ec2.internal + rank : 62 (local_rank: 6) + exitcode : 1 (pid: 883146) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-236.ec2.internal + rank : 63 (local_rank: 7) + exitcode : 1 (pid: 883147) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-236.ec2.internal + rank : 56 (local_rank: 0) + exitcode : 1 (pid: 883140) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + sys.exit(main()) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + run(args) + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-226.ec2.internal + rank : 49 (local_rank: 1) + exitcode : 1 (pid: 3119068) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-226.ec2.internal + rank : 50 (local_rank: 2) + exitcode : 1 (pid: 3119069) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-226.ec2.internal + rank : 51 (local_rank: 3) + exitcode : 1 (pid: 3119070) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-226.ec2.internal + rank : 52 (local_rank: 4) + exitcode : 1 (pid: 3119071) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-226.ec2.internal + rank : 53 (local_rank: 5) + exitcode : 1 (pid: 3119072) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-226.ec2.internal + rank : 54 (local_rank: 6) + exitcode : 1 (pid: 3119073) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-226.ec2.internal + rank : 55 (local_rank: 7) + exitcode : 1 (pid: 3119074) + error_file: + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_18:42:52 + host : ip-26-0-163-226.ec2.internal + rank : 48 (local_rank: 0) + exitcode : 1 (pid: 3119067) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-14.ec2.internal + rank : 25 (local_rank: 1) + exitcode : 1 (pid: 2461731) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-14.ec2.internal + rank : 26 (local_rank: 2) + exitcode : 1 (pid: 2461732) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-14.ec2.internal + rank : 27 (local_rank: 3) + exitcode : 1 (pid: 2461733) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-14.ec2.internal + rank : 28 (local_rank: 4) + exitcode : 1 (pid: 2461734) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-14.ec2.internal + rank : 29 (local_rank: 5) + exitcode : 1 (pid: 2461735) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-14.ec2.internal + rank : 30 (local_rank: 6) + exitcode : 1 (pid: 2461736) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-14.ec2.internal + rank : 31 (local_rank: 7) + exitcode : 1 (pid: 2461737) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-27_18:42:52 + host : ip-26-0-162-14.ec2.internal + rank : 24 (local_rank: 0) + exitcode : 1 (pid: 2461730) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-161-221: task 2: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13803749.0 +slurmstepd: error: *** STEP 13803749.0 ON ip-26-0-161-153 CANCELLED AT 2024-12-27T18:42:52 *** +[2024-12-27 18:42:52,527] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-27 18:42:52,527] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1600806 closing signal SIGTERM +[2024-12-27 18:42:52,527] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1600809 closing signal SIGTERM +[2024-12-27 18:42:52,527] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1600810 closing signal SIGTERM +srun: error: ip-26-0-161-178: task 1: Exited with exit code 1 +srun: error: ip-26-0-163-226: task 6: Exited with exit code 1 +srun: error: ip-26-0-163-236: task 7: Terminated +srun: error: ip-26-0-162-46: task 4: Terminated +srun: error: ip-26-0-162-14: task 3: Terminated +srun: error: ip-26-0-161-153: task 0: Terminated +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1600731 got signal: 15 +srun: error: ip-26-0-163-220: task 5: Exited with exit code 1 +srun: Force Terminated StepId=13803749.0 diff --git a/logs/13848169-bench_3.57G_dp2_tp1_pp8_acc32_mbs4_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13848169-bench_3.57G_dp2_tp1_pp8_acc32_mbs4_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..71f9add0bc5c24adacd14427ac33cf65f4d32893 --- /dev/null +++ b/logs/13848169-bench_3.57G_dp2_tp1_pp8_acc32_mbs4_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,10476 @@ ++ '[' -z 13848169 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-170-[143,160]' ++ export 'NODELIST=ip-26-0-170-143 +ip-26-0-170-160' ++ NODELIST='ip-26-0-170-143 +ip-26-0-170-160' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-170-[143,160]' ++ export MASTER_NODE=ip-26-0-170-143 ++ MASTER_NODE=ip-26-0-170-143 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-170-143' +Master node: ip-26-0-170-143 ++ echo 'All nodes: ip-26-0-170-143 +ip-26-0-170-160' +All nodes: ip-26-0-170-143 +ip-26-0-170-160 ++ echo 'World size: 16' +World size: 16 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13848169 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-170-143:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_3.57G_dp2_tp1_pp8_acc32_mbs4_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-28 00:52:52,141] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 00:52:52,141] torch.distributed.run: [WARNING] +[2024-12-28 00:52:52,141] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 00:52:52,141] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 00:52:52,141] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 00:52:52,239] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 00:52:52,239] torch.distributed.run: [WARNING] +[2024-12-28 00:52:52,239] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 00:52:52,239] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 00:52:52,239] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-170-143:2167363:2167363 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-143:2167363:2167363 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2167363:2167363 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2167363:2167363 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2167363:2167363 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:65609:65609 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:65604:65604 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:65609:65609 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-160:65604:65604 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-160:65609:65609 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:65604:65604 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:65609:65609 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:65609:65609 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:65604:65604 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:65604:65604 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2167365:2167365 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2167366:2167366 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2167368:2167368 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2167364:2167364 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2167369:2167369 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2167366:2167366 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-143:2167365:2167365 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-143:2167368:2167368 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-143:2167364:2167364 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-143:2167369:2167369 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-143:2167367:2167367 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2167367:2167367 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-143:2167370:2167370 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-143:2167370:2167370 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-143:2167366:2167366 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2167365:2167365 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2167364:2167364 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2167368:2167368 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2167369:2167369 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2167367:2167367 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2167365:2167365 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2167366:2167366 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2167366:2167366 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2167365:2167365 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2167364:2167364 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2167364:2167364 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2167368:2167368 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2167368:2167368 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2167369:2167369 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2167369:2167369 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2167370:2167370 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.170.143<0> +ip-26-0-170-143:2167367:2167367 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2167367:2167367 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-143:2167370:2167370 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-143:2167370:2167370 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:65610:65610 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:65606:65606 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:65605:65605 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:65608:65608 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:65610:65610 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-160:65605:65605 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-160:65606:65606 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-160:65611:65611 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:65608:65608 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-160:65611:65611 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-160:65607:65607 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-170-160:65607:65607 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-170-160:65610:65610 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:65605:65605 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:65608:65608 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:65606:65606 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:65610:65610 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:65605:65605 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:65605:65605 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:65610:65610 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:65611:65611 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:65608:65608 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:65608:65608 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:65607:65607 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.170.160<0> +ip-26-0-170-160:65606:65606 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:65606:65606 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:65607:65607 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:65607:65607 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:65611:65611 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-170-160:65611:65611 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:65611:65768 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:65611:65768 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/205 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:65608:65767 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:65608:65767 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:65604:65764 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:65604:65764 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:65609:65763 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:65609:65763 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:65610:65766 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:65610:65766 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:65606:65765 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:65606:65765 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:65607:65770 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:65607:65770 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-160:65605:65769 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-160:65605:65769 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO comm 0x9a40ba0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO comm 0x9fc7730 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO comm 0x9726610 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO comm 0x904ea10 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO comm 0x909f8b0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO comm 0x8523ee0 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-160:65604:65764 [0] NCCL INFO comm 0x907d770 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-160:65605:65769 [1] NCCL INFO comm 0x9c7a760 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO comm 0x8b7bbe0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-160:65611:65768 [7] NCCL INFO comm 0x8d39c10 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-160:65606:65765 [2] NCCL INFO comm 0x93b3a90 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-160:65607:65770 [3] NCCL INFO comm 0x9cb8b50 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-160:65610:65766 [6] NCCL INFO comm 0x9739f30 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-160:65609:65763 [5] NCCL INFO comm 0x8cbe910 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-160:65608:65767 [4] NCCL INFO comm 0x96abca0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO comm 0x894a7f0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf2c38145403119cf - Init START +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65605:65769 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65607:65770 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65606:65765 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-170-160:65610:65766 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65610:65766 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-170-160:65606:65765 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:65606:65765 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-170-160:65609:65763 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65609:65763 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-170-160:65605:65769 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:65605:65769 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-170-160:65607:65770 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:65607:65770 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-170-160:65611:65768 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65611:65768 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-170-160:65604:65764 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:65604:65764 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-170-160:65608:65767 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65608:65767 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-170-160:65607:65770 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->3 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/3/-1->11->-1 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 +ip-26-0-170-160:65607:65770 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65608:65767 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->4 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/4/-1->12->-1 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 +ip-26-0-170-160:65608:65767 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-160:65609:65763 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->5 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/5/-1->13->-1 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 +ip-26-0-170-160:65609:65763 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65611:65768 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->7 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/7/-1->15->-1 +ip-26-0-170-160:65611:65768 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-160:65610:65766 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->6 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/6/-1->14->-1 [15] -1/-1/-1->14->13 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/14/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-160:65610:65766 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65604:65764 [0] NCCL INFO Trees [0] 9/-1/-1->8->0 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/0/-1->8->-1 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/15/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-160:65604:65764 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65606:65765 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->2 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/2/-1->10->-1 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 +ip-26-0-170-160:65605:65769 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->1 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/1/-1->9->-1 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/13/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 01/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 1 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-160:65606:65765 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65605:65769 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 2 1 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 03/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 3 2 1 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 04/16 : 0 7 6 5 12 11 10 9 8 15 14 13 4 3 2 1 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/12/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 05/16 : 0 4 7 6 13 11 10 9 8 12 15 14 5 3 2 1 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 06/16 : 0 5 4 7 14 11 10 9 8 13 12 15 6 3 2 1 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/10/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 07/16 : 0 6 5 4 15 11 10 9 8 14 13 12 7 3 2 1 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/9/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 09/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 1 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 2 1 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 11/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 3 2 1 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 12/16 : 0 7 6 5 12 11 10 9 8 15 14 13 4 3 2 1 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 13/16 : 0 4 7 6 13 11 10 9 8 12 15 14 5 3 2 1 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 14/16 : 0 5 4 7 14 11 10 9 8 13 12 15 6 3 2 1 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/11/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 15/16 : 0 6 5 4 15 11 10 9 8 14 13 12 7 3 2 1 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Trees [0] 1/8/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 05/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 06/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 13/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 14/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 05/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 06/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 13/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 14/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 08/0 : 15[7] -> 0[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 08/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 05/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 13/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 06/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 03/0 : 12[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 11/0 : 12[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 14/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 06/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 07/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 14/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 15/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 07/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 02/0 : 11[3] -> 2[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 10/0 : 11[3] -> 2[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 15/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 02/0 : 11[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 10/0 : 11[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 06/0 : 15[7] -> 6[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 01/0 : 10[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 09/0 : 10[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 01/0 : 10[2] -> 1[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 09/0 : 10[2] -> 1[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 03/0 : 12[4] -> 3[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 11/0 : 12[4] -> 3[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 04/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 12/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 14/0 : 15[7] -> 6[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 05/0 : 14[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 13/0 : 14[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 05/0 : 14[6] -> 5[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 13/0 : 14[6] -> 5[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 04/0 : 13[5] -> 4[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 12/0 : 13[5] -> 4[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 06/0 : 15[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 14/0 : 15[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 05/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 06/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 13/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 14/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 07/0 : 12[4] -> 7[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 15/0 : 12[4] -> 7[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 09/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 10/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 06/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 05/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 13/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 14/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 07/0 : 12[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 15/0 : 12[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 07/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 15/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 12/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 11/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Connected all rings +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Connected all rings +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Connected all rings +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Connected all rings +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Connected all rings +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Connected all rings +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Connected all rings +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Connected all rings +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Connected all rings +ip-26-0-170-160:65611:65768 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Connected all rings +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Connected all rings +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Connected all rings +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65607:65770 [3] NCCL INFO Connected all trees +ip-26-0-170-160:65607:65770 [3] NCCL INFO NVLS comm 0x9cb8b50 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65606:65765 [2] NCCL INFO Connected all trees +ip-26-0-170-160:65606:65765 [2] NCCL INFO NVLS comm 0x93b3a90 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65608:65767 [4] NCCL INFO Connected all trees +ip-26-0-170-160:65608:65767 [4] NCCL INFO NVLS comm 0x96abca0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65609:65763 [5] NCCL INFO Connected all trees +ip-26-0-170-160:65609:65763 [5] NCCL INFO NVLS comm 0x8cbe910 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65610:65766 [6] NCCL INFO Connected all trees +ip-26-0-170-160:65604:65764 [0] NCCL INFO Connected all trees +ip-26-0-170-160:65610:65766 [6] NCCL INFO NVLS comm 0x9739f30 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65604:65764 [0] NCCL INFO NVLS comm 0x907d770 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NVLS comm 0x909f8b0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NVLS comm 0x9fc7730 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NVLS comm 0x9726610 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NVLS comm 0x894a7f0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NVLS comm 0x904ea10 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65611:65768 [7] NCCL INFO Connected all trees +ip-26-0-170-160:65611:65768 [7] NCCL INFO NVLS comm 0x8d39c10 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65605:65769 [1] NCCL INFO Connected all trees +ip-26-0-170-160:65605:65769 [1] NCCL INFO NVLS comm 0x9c7a760 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NVLS comm 0x9a40ba0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NVLS comm 0x8523ee0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NVLS comm 0x8b7bbe0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65768 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65605:65769 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65609:65763 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:65767 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65765 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65770 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65764 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65766 [6] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65610:65766 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65610:65766 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65610:65766 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65607:65770 [3] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65607:65770 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65607:65770 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65607:65770 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65611:65768 [7] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65611:65768 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65611:65768 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65611:65768 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65605:65769 [1] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65605:65769 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65605:65769 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65605:65769 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65604:65764 [0] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65604:65764 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65604:65764 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65604:65764 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65608:65767 [4] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65608:65767 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65608:65767 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65608:65767 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65606:65765 [2] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65606:65765 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65606:65765 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65606:65765 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65609:65763 [5] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65609:65763 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65609:65763 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65609:65763 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65608:65767 [4] NCCL INFO comm 0x96abca0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-160:65606:65765 [2] NCCL INFO comm 0x93b3a90 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-160:65604:65764 [0] NCCL INFO comm 0x907d770 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-160:65610:65766 [6] NCCL INFO comm 0x9739f30 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-160:65609:65763 [5] NCCL INFO comm 0x8cbe910 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-160:65611:65768 [7] NCCL INFO comm 0x8d39c10 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-160:65605:65769 [1] NCCL INFO comm 0x9c7a760 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-160:65607:65770 [3] NCCL INFO comm 0x9cb8b50 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-143:2167368:2167524 [5] NCCL INFO comm 0x9726610 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-143:2167366:2167527 [3] NCCL INFO comm 0x9fc7730 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-143:2167364:2167526 [1] NCCL INFO comm 0x8b7bbe0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-143:2167367:2167529 [4] NCCL INFO comm 0x904ea10 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-143:2167363:2167523 [0] NCCL INFO comm 0x894a7f0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-143:2167365:2167530 [2] NCCL INFO comm 0x909f8b0 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-143:2167370:2167528 [7] NCCL INFO comm 0x9a40ba0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-143:2167369:2167525 [6] NCCL INFO comm 0x8523ee0 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf2c38145403119cf - Init COMPLETE +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:65609:65933 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:65610:65932 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:65608:65931 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:65605:65937 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:65607:65935 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:65604:65934 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:65606:65938 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:65611:65936 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO comm 0x8b808b0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO comm 0x8dae7c0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO comm 0xa1f9900 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-160:65610:65932 [6] NCCL INFO comm 0x996c9a0 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-160:65611:65936 [7] NCCL INFO comm 0x8f6ce40 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-160:65609:65933 [5] NCCL INFO comm 0x8ef1de0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-160:65608:65931 [4] NCCL INFO comm 0x98de420 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO comm 0x9958ea0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO comm 0x92d2300 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO comm 0x8757360 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO comm 0x9280870 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO comm 0x9c73880 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-160:65607:65935 [3] NCCL INFO comm 0x9eec400 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-160:65606:65938 [2] NCCL INFO comm 0x95e6bc0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-160:65605:65937 [1] NCCL INFO comm 0x9eae360 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-160:65604:65934 [0] NCCL INFO comm 0x92b0020 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7f1cc2d083d835f8 - Init START +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-160:65610:65932 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65611:65936 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65607:65935 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65605:65937 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65606:65938 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65609:65933 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65604:65934 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65608:65931 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65606:65938 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:65606:65938 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-170-160:65609:65933 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65609:65933 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-170-160:65611:65936 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65611:65936 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-170-160:65608:65931 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65608:65931 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-170-160:65610:65932 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65610:65932 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-170-160:65605:65937 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:65605:65937 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-170-160:65607:65935 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:65607:65935 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-170-160:65604:65934 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:65604:65934 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-170-160:65607:65935 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->3 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/3/-1->11->-1 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 +ip-26-0-170-160:65607:65935 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65608:65931 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->4 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/4/-1->12->-1 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 +ip-26-0-170-160:65608:65931 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65609:65933 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->5 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/5/-1->13->-1 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 +ip-26-0-170-160:65611:65936 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->7 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/7/-1->15->-1 +ip-26-0-170-160:65609:65933 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65611:65936 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-160:65610:65932 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->6 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/6/-1->14->-1 [15] -1/-1/-1->14->13 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-160:65610:65932 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 01/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 1 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 2 1 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 03/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 3 2 1 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/9/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 04/16 : 0 7 6 5 12 11 10 9 8 15 14 13 4 3 2 1 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 05/16 : 0 4 7 6 13 11 10 9 8 12 15 14 5 3 2 1 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 06/16 : 0 5 4 7 14 11 10 9 8 13 12 15 6 3 2 1 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 07/16 : 0 6 5 4 15 11 10 9 8 14 13 12 7 3 2 1 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/10/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 09/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 1 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 2 1 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 11/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 3 2 1 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 12/16 : 0 7 6 5 12 11 10 9 8 15 14 13 4 3 2 1 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 13/16 : 0 4 7 6 13 11 10 9 8 12 15 14 5 3 2 1 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 14/16 : 0 5 4 7 14 11 10 9 8 13 12 15 6 3 2 1 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 15/16 : 0 6 5 4 15 11 10 9 8 14 13 12 7 3 2 1 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-160:65606:65938 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->2 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/2/-1->10->-1 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 +ip-26-0-170-160:65606:65938 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65604:65934 [0] NCCL INFO Trees [0] 9/-1/-1->8->0 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/0/-1->8->-1 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 +ip-26-0-170-160:65604:65934 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Trees [0] 1/8/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-160:65605:65937 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->1 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/1/-1->9->-1 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 +ip-26-0-170-160:65605:65937 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/13/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/12/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/15/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/11/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/14/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 05/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 06/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 13/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 14/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 05/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 06/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 13/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 14/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 08/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 03/0 : 12[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 11/0 : 12[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 02/0 : 11[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 10/0 : 11[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 08/0 : 15[7] -> 0[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 05/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 13/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 02/0 : 11[3] -> 2[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 10/0 : 11[3] -> 2[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 06/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 06/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 14/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 14/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 07/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 07/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 15/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 15/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 06/0 : 15[7] -> 6[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 01/0 : 10[2] -> 1[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 01/0 : 10[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 09/0 : 10[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 03/0 : 12[4] -> 3[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 11/0 : 12[4] -> 3[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 14/0 : 15[7] -> 6[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 05/0 : 14[6] -> 5[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 13/0 : 14[6] -> 5[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 04/0 : 13[5] -> 4[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 12/0 : 13[5] -> 4[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 04/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 12/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 05/0 : 14[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 13/0 : 14[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 06/0 : 15[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 14/0 : 15[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 06/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 06/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 05/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 14/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 13/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 05/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 14/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 13/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 09/0 : 10[2] -> 1[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 10/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 09/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 07/0 : 12[4] -> 7[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 07/0 : 12[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 15/0 : 12[4] -> 7[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 15/0 : 12[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 07/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 15/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 12/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 11/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Connected all rings +ip-26-0-170-160:65606:65938 [2] NCCL INFO Connected all rings +ip-26-0-170-160:65608:65931 [4] NCCL INFO Connected all rings +ip-26-0-170-160:65610:65932 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Connected all rings +ip-26-0-170-160:65611:65936 [7] NCCL INFO Connected all rings +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Connected all rings +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Connected all rings +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Connected all rings +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Connected all rings +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Connected all rings +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Connected all rings +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Connected all rings +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65606:65938 [2] NCCL INFO Connected all trees +ip-26-0-170-160:65606:65938 [2] NCCL INFO NVLS comm 0x95e6bc0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65610:65932 [6] NCCL INFO Connected all trees +ip-26-0-170-160:65610:65932 [6] NCCL INFO NVLS comm 0x996c9a0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65604:65934 [0] NCCL INFO Connected all trees +ip-26-0-170-160:65604:65934 [0] NCCL INFO NVLS comm 0x92b0020 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Connected all trees +ip-26-0-170-160:65607:65935 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NVLS comm 0x92d2300 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65607:65935 [3] NCCL INFO NVLS comm 0x9eec400 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65611:65936 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Connected all trees +ip-26-0-170-160:65611:65936 [7] NCCL INFO NVLS comm 0x8f6ce40 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NVLS comm 0x8757360 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65605:65937 [1] NCCL INFO Connected all trees +ip-26-0-170-160:65605:65937 [1] NCCL INFO NVLS comm 0x9eae360 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NVLS comm 0x8b808b0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NVLS comm 0x9c73880 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Connected all trees +ip-26-0-170-160:65609:65933 [5] NCCL INFO Connected all trees +ip-26-0-170-160:65608:65931 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NVLS comm 0x8dae7c0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65609:65933 [5] NCCL INFO NVLS comm 0x8ef1de0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65608:65931 [4] NCCL INFO NVLS comm 0x98de420 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NVLS comm 0xa1f9900 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NVLS comm 0x9958ea0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NVLS comm 0x9280870 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:65932 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:65937 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65604:65934 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65608:65931 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:65936 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65609:65933 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65606:65938 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65607:65935 [3] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65607:65935 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65607:65935 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65607:65935 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65605:65937 [1] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65605:65937 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65605:65937 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65605:65937 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65611:65936 [7] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65611:65936 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65611:65936 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65611:65936 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65606:65938 [2] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65606:65938 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65606:65938 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65606:65938 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65610:65932 [6] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65610:65932 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65610:65932 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65610:65932 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65608:65931 [4] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65608:65931 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65608:65931 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65608:65931 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65609:65933 [5] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65609:65933 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65609:65933 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65609:65933 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65604:65934 [0] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65604:65934 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65604:65934 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-160:65604:65934 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65610:65932 [6] NCCL INFO comm 0x996c9a0 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-160:65606:65938 [2] NCCL INFO comm 0x95e6bc0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-160:65608:65931 [4] NCCL INFO comm 0x98de420 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-160:65604:65934 [0] NCCL INFO comm 0x92b0020 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-160:65609:65933 [5] NCCL INFO comm 0x8ef1de0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-160:65611:65936 [7] NCCL INFO comm 0x8f6ce40 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-160:65605:65937 [1] NCCL INFO comm 0x9eae360 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-160:65607:65935 [3] NCCL INFO comm 0x9eec400 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-143:2167365:2167696 [2] NCCL INFO comm 0x92d2300 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-143:2167368:2167697 [5] NCCL INFO comm 0x9958ea0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-143:2167364:2167693 [1] NCCL INFO comm 0x8dae7c0 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-143:2167366:2167694 [3] NCCL INFO comm 0xa1f9900 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-143:2167367:2167698 [4] NCCL INFO comm 0x9280870 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-143:2167369:2167695 [6] NCCL INFO comm 0x8757360 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-143:2167363:2167692 [0] NCCL INFO comm 0x8b808b0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +ip-26-0-170-143:2167370:2167699 [7] NCCL INFO comm 0x9c73880 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7f1cc2d083d835f8 - Init COMPLETE +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:65608:65968 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:65604:65966 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:65609:65967 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:65610:65969 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:65611:65971 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:65606:65970 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:65607:65973 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO comm 0x996f290 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc7b2c716f5097409 - Init START +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO comm 0x876d130 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc7b2c716f5097409 - Init START +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO comm 0x92971c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc7b2c716f5097409 - Init START +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO comm 0x9c8aca0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc7b2c716f5097409 - Init START +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO comm 0xa210740 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc7b2c716f5097409 - Init START +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO comm 0x92e93d0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc7b2c716f5097409 - Init START +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO comm 0x8dc5780 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc7b2c716f5097409 - Init START +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO comm 0x8b98680 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc7b2c716f5097409 - Init START +ip-26-0-170-160:65605:65972 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-160:65605:65972 [1] NCCL INFO comm 0x9ec4b00 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5978c19f744b735a - Init START +ip-26-0-170-160:65611:65971 [7] NCCL INFO comm 0x8f84740 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5978c19f744b735a - Init START +ip-26-0-170-160:65604:65966 [0] NCCL INFO comm 0x92c9e10 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5978c19f744b735a - Init START +ip-26-0-170-160:65610:65969 [6] NCCL INFO comm 0x9983380 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5978c19f744b735a - Init START +ip-26-0-170-160:65609:65967 [5] NCCL INFO comm 0x8f07760 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5978c19f744b735a - Init START +ip-26-0-170-160:65608:65968 [4] NCCL INFO comm 0x98f4ea0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5978c19f744b735a - Init START +ip-26-0-170-160:65607:65973 [3] NCCL INFO comm 0x9f031c0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5978c19f744b735a - Init START +ip-26-0-170-160:65606:65970 [2] NCCL INFO comm 0x95fc760 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5978c19f744b735a - Init START +ip-26-0-170-160:65611:65971 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65605:65972 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65604:65966 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65610:65969 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65609:65967 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65608:65968 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65606:65970 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65607:65973 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:65604:65966 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-170-160:65608:65968 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65608:65968 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-170-160:65605:65972 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:65605:65972 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-170-160:65607:65973 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:65607:65973 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65610:65969 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65610:65969 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:65606:65970 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65611:65971 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65609:65967 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-170-160:65609:65967 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-170-160:65609:65967 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65610:65969 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-170-160:65610:65969 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65608:65968 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65606:65970 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-170-160:65608:65968 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65611:65971 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-170-160:65607:65973 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65606:65970 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65611:65971 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65605:65972 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65607:65973 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65605:65972 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-170-160:65604:65966 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-170-160:65604:65966 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Connected all rings +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Connected all rings +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Connected all rings +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Connected all rings +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Connected all rings +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65610:65969 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO NVLS comm 0x9c8aca0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65608:65968 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65609:65967 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO Connected all trees +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65606:65970 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-170-160:65611:65971 [7] NCCL INFO NVLS comm 0x8f84740 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO NVLS comm 0x8b98680 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65607:65973 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65605:65972 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:65966 [0] NCCL INFO Connected all trees +ip-26-0-170-160:65604:65966 [0] NCCL INFO NVLS comm 0x92c9e10 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65609:65967 [5] NCCL INFO Connected all trees +ip-26-0-170-160:65609:65967 [5] NCCL INFO NVLS comm 0x8f07760 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO NVLS comm 0x996f290 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO NVLS comm 0xa210740 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO NVLS comm 0x92971c0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO NVLS comm 0x8dc5780 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO NVLS comm 0x876d130 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO NVLS comm 0x92e93d0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65610:65969 [6] NCCL INFO Connected all trees +ip-26-0-170-160:65610:65969 [6] NCCL INFO NVLS comm 0x9983380 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65608:65968 [4] NCCL INFO Connected all trees +ip-26-0-170-160:65608:65968 [4] NCCL INFO NVLS comm 0x98f4ea0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65606:65970 [2] NCCL INFO Connected all trees +ip-26-0-170-160:65606:65970 [2] NCCL INFO NVLS comm 0x95fc760 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65605:65972 [1] NCCL INFO Connected all trees +ip-26-0-170-160:65605:65972 [1] NCCL INFO NVLS comm 0x9ec4b00 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-160:65607:65973 [3] NCCL INFO Connected all trees +ip-26-0-170-160:65607:65973 [3] NCCL INFO NVLS comm 0x9f031c0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167366:2167732 [3] NCCL INFO comm 0xa210740 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc7b2c716f5097409 - Init COMPLETE +ip-26-0-170-143:2167364:2167734 [1] NCCL INFO comm 0x8dc5780 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc7b2c716f5097409 - Init COMPLETE +ip-26-0-170-143:2167370:2167733 [7] NCCL INFO comm 0x9c8aca0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc7b2c716f5097409 - Init COMPLETE +ip-26-0-170-143:2167368:2167730 [5] NCCL INFO comm 0x996f290 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc7b2c716f5097409 - Init COMPLETE +ip-26-0-170-143:2167363:2167728 [0] NCCL INFO comm 0x8b98680 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc7b2c716f5097409 - Init COMPLETE +ip-26-0-170-143:2167367:2167729 [4] NCCL INFO comm 0x92971c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc7b2c716f5097409 - Init COMPLETE +ip-26-0-170-143:2167369:2167731 [6] NCCL INFO comm 0x876d130 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc7b2c716f5097409 - Init COMPLETE +ip-26-0-170-143:2167365:2167735 [2] NCCL INFO comm 0x92e93d0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc7b2c716f5097409 - Init COMPLETE +ip-26-0-170-160:65610:65969 [6] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65610:65969 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65608:65968 [4] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65610:65969 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65610:65969 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65608:65968 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65608:65968 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65608:65968 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65606:65970 [2] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65606:65970 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65606:65970 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65606:65970 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65609:65967 [5] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65609:65967 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65609:65967 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65609:65967 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65604:65966 [0] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65604:65966 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65607:65973 [3] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65607:65973 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65607:65973 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65605:65972 [1] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65611:65971 [7] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65607:65973 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65604:65966 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65611:65971 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65604:65966 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65605:65972 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65611:65971 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65611:65971 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65605:65972 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65605:65972 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65610:65969 [6] NCCL INFO comm 0x9983380 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5978c19f744b735a - Init COMPLETE +ip-26-0-170-160:65608:65968 [4] NCCL INFO comm 0x98f4ea0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5978c19f744b735a - Init COMPLETE +ip-26-0-170-160:65606:65970 [2] NCCL INFO comm 0x95fc760 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5978c19f744b735a - Init COMPLETE +ip-26-0-170-160:65607:65973 [3] NCCL INFO comm 0x9f031c0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5978c19f744b735a - Init COMPLETE +ip-26-0-170-160:65609:65967 [5] NCCL INFO comm 0x8f07760 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5978c19f744b735a - Init COMPLETE +ip-26-0-170-160:65604:65966 [0] NCCL INFO comm 0x92c9e10 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5978c19f744b735a - Init COMPLETE +ip-26-0-170-160:65605:65972 [1] NCCL INFO comm 0x9ec4b00 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5978c19f744b735a - Init COMPLETE +ip-26-0-170-160:65611:65971 [7] NCCL INFO comm 0x8f84740 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5978c19f744b735a - Init COMPLETE +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Config: +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Config(general=GeneralArgs(project='debug', +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: run='3.57G_dp2_tp1_pp8_acc32_mbs4_seq4096_zero1_tpmodeRED_vocab131k', +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: seed=42, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: step=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: consumed_train_samples=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: ignore_sanity_checks=True), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: parallelism=ParallelismArgs(dp=2, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pp=8, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp=1, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pp_engine=, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp_mode=, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp_linear_async_communication=True, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: recompute_layer=False, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tp_recompute_allgather=True, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: expert_parallel_size=1), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: eos_token_id=0, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_act='silu', +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_size=3072, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: initializer_range=0.02, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: intermediate_size=8192, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: is_llama_config=True, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: max_position_embeddings=4096, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_attention_heads=32, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_hidden_layers=28, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_key_value_heads=32, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pad_token_id=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pretraining_tp=1, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rms_norm_eps=1e-05, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_scaling=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_theta=10000.0, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_interleaved=False, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tie_word_embeddings=True, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: use_cache=True, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: vocab_size=131072), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: init_method=RandomInit(std=0.02), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: dtype=torch.bfloat16, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: make_vocab_size_divisible_by=1, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: ddp_bucket_cap_mb=25), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokenizer_revision=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokenizer_max_length=None), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: checkpoint_interval=10000, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: save_initial_state=False, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: save_final_state=False, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: resume_checkpoint_path=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: checkpoints_path_is_shared_file_system=False), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: logging=LoggingArgs(log_level='info', +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: log_level_replica='info', +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: iteration_step_info_interval=1), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tokens=TokensArgs(sequence_length=4096, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: train_steps=100, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: micro_batch_size=4, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: batch_accumulation_per_replica=32, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: val_check_interval=100, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: limit_val_batches=0, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: limit_test_batches=0), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: adam_beta1=0.9, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: adam_beta2=0.95, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: torch_adam_is_fused=True, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: name='adamW'), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: zero_stage=1, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: weight_decay=0.01, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: clip_grad=1.0, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: accumulate_grad_in_fp32=True, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_warmup_steps=2, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_warmup_style='linear', +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_decay_style='cosine', +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_decay_steps=13, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lr_decay_starting_step=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: min_decay_lr=1e-05)), +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: start_training_step=1, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: data=DataArgs(dataset=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: seed=42, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_loading_workers=1))], +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: profiler=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: lighteval=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: s3_upload=None) +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Model Config: +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: LlamaConfig(bos_token_id=0, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: eos_token_id=0, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_act='silu', +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: hidden_size=3072, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: initializer_range=0.02, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: intermediate_size=8192, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: is_llama_config=True, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: max_position_embeddings=4096, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_attention_heads=32, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_hidden_layers=28, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: num_key_value_heads=32, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pad_token_id=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: pretraining_tp=1, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rms_norm_eps=1e-05, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_scaling=None, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_theta=10000.0, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: rope_interleaved=False, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: tie_word_embeddings=True, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: use_cache=True, +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: vocab_size=131072) +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Building model.. +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Initialize RoPE Theta = 10000.0 +12/28/2024 00:53:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:65606:66007 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:65605:66011 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:65608:66009 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:65607:66012 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:65609:66010 [5] NCCL INFO comm 0xab80eb0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1376ee165f4632f2 - Init START +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO comm 0xaf11c30 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc75a92f5e2ae260f - Init START +ip-26-0-170-160:65605:66011 [1] NCCL INFO comm 0xbb3e580 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbad458b8b5bc1873 - Init START +ip-26-0-170-160:65606:66007 [2] NCCL INFO comm 0xb2748b0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xaa08efda9571de89 - Init START +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65609:66010 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65608:66009 [4] NCCL INFO comm 0xb56f7f0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfeb44020e5f8e3f9 - Init START +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO comm 0xb5e9e00 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3b1d5eee952fd9e - Init START +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO comm 0xa3e7a70 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa675181babd8e5f9 - Init START +ip-26-0-170-160:65605:66011 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65607:66012 [3] NCCL INFO comm 0xbb7b6b0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf1945d70782c370f - Init START +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:65606:66007 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO comm 0xbe89ef0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x28a85d9b77c9c6a2 - Init START +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65608:66009 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65607:66012 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO comm 0xaf62df0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2b95f9619d5a5d7f - Init START +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO comm 0xb904800 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x446c8f7ca4819f5d - Init START +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:65604:66016 [0] NCCL INFO comm 0xaf43db0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xae842a852b97c763 - Init START +ip-26-0-170-160:65604:66016 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-170-160:65610:66022 [6] NCCL INFO comm 0xaba03a0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4af62f2ca07147cd - Init START +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO comm 0xa9d8f30 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5c8080bd8a04bc67 - Init START +ip-26-0-170-160:65610:66022 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-160:65611:66024 [7] NCCL INFO comm 0xa19eca0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc03b0c366dee6e83 - Init START +ip-26-0-170-160:65611:66024 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO comm 0xac03d20 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x339237567ae63c58 - Init START +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:65604:66016 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65604:66016 [0] NCCL INFO Connected all rings +ip-26-0-170-160:65604:66016 [0] NCCL INFO Connected all trees +ip-26-0-170-160:65604:66016 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65609:66010 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:65609:66010 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65609:66010 [5] NCCL INFO Connected all rings +ip-26-0-170-160:65609:66010 [5] NCCL INFO Connected all trees +ip-26-0-170-160:65609:66010 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65604:66016 [0] NCCL INFO comm 0xaf43db0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xae842a852b97c763 - Init COMPLETE +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65605:66011 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:65605:66011 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65605:66011 [1] NCCL INFO Connected all rings +ip-26-0-170-160:65605:66011 [1] NCCL INFO Connected all trees +ip-26-0-170-160:65605:66011 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65610:66022 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:65610:66022 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65610:66022 [6] NCCL INFO Connected all rings +ip-26-0-170-160:65610:66022 [6] NCCL INFO Connected all trees +ip-26-0-170-160:65610:66022 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65609:66010 [5] NCCL INFO comm 0xab80eb0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1376ee165f4632f2 - Init COMPLETE +ip-26-0-170-160:65606:66007 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:65606:66007 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65606:66007 [2] NCCL INFO Connected all rings +ip-26-0-170-160:65606:66007 [2] NCCL INFO Connected all trees +ip-26-0-170-160:65606:66007 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167783 [0] NCCL INFO comm 0xa9d8f30 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5c8080bd8a04bc67 - Init COMPLETE +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65611:66024 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:65611:66024 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65611:66024 [7] NCCL INFO Connected all rings +ip-26-0-170-160:65611:66024 [7] NCCL INFO Connected all trees +ip-26-0-170-160:65611:66024 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65610:66022 [6] NCCL INFO comm 0xaba03a0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4af62f2ca07147cd - Init COMPLETE +ip-26-0-170-160:65605:66011 [1] NCCL INFO comm 0xbb3e580 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbad458b8b5bc1873 - Init COMPLETE +ip-26-0-170-160:65607:66012 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:65607:66012 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65607:66012 [3] NCCL INFO Connected all rings +ip-26-0-170-160:65607:66012 [3] NCCL INFO Connected all trees +ip-26-0-170-160:65607:66012 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65606:66007 [2] NCCL INFO comm 0xb2748b0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xaa08efda9571de89 - Init COMPLETE +ip-26-0-170-160:65608:66009 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-160:65608:66009 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65608:66009 [4] NCCL INFO Connected all rings +ip-26-0-170-160:65608:66009 [4] NCCL INFO Connected all trees +ip-26-0-170-160:65608:66009 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65611:66024 [7] NCCL INFO comm 0xa19eca0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc03b0c366dee6e83 - Init COMPLETE +ip-26-0-170-160:65607:66012 [3] NCCL INFO comm 0xbb7b6b0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf1945d70782c370f - Init COMPLETE +ip-26-0-170-143:2167369:2167767 [6] NCCL INFO comm 0xa3e7a70 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa675181babd8e5f9 - Init COMPLETE +ip-26-0-170-143:2167366:2167771 [3] NCCL INFO comm 0xbe89ef0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x28a85d9b77c9c6a2 - Init COMPLETE +ip-26-0-170-160:65608:66009 [4] NCCL INFO comm 0xb56f7f0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfeb44020e5f8e3f9 - Init COMPLETE +ip-26-0-170-143:2167368:2167770 [5] NCCL INFO comm 0xb5e9e00 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3b1d5eee952fd9e - Init COMPLETE +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167365:2167772 [2] NCCL INFO comm 0xaf62df0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2b95f9619d5a5d7f - Init COMPLETE +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167367:2167768 [4] NCCL INFO comm 0xaf11c30 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc75a92f5e2ae260f - Init COMPLETE +ip-26-0-170-143:2167364:2167785 [1] NCCL INFO comm 0xac03d20 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x339237567ae63c58 - Init COMPLETE +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:65610:66044 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:65608:66042 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:65606:66043 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:65604:66045 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167370:2167774 [7] NCCL INFO comm 0xb904800 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x446c8f7ca4819f5d - Init COMPLETE +ip-26-0-170-160:65610:66044 [6] NCCL INFO comm 0xaba3320 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb259cfe95f7000d5 - Init START +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO comm 0xa3ea720 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb259cfe95f7000d5 - Init START +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO comm 0xaf148e0 rank 2 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb259cfe95f7000d5 - Init START +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO comm 0xaf65aa0 rank 1 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb259cfe95f7000d5 - Init START +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO comm 0xa9dbbe0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb259cfe95f7000d5 - Init START +ip-26-0-170-160:65608:66042 [4] NCCL INFO comm 0xb5724a0 rank 6 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb259cfe95f7000d5 - Init START +ip-26-0-170-160:65606:66043 [2] NCCL INFO comm 0xb277560 rank 5 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb259cfe95f7000d5 - Init START +ip-26-0-170-160:65604:66045 [0] NCCL INFO comm 0xaf46a60 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb259cfe95f7000d5 - Init START +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-160:65608:66042 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65610:66044 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65606:66043 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65604:66045 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:65607:66047 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:65609:66046 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:65611:66048 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:65605:66049 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO comm 0xb5ecab0 rank 2 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7d517d840603fb25 - Init START +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO comm 0xbe8cba0 rank 1 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7d517d840603fb25 - Init START +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO comm 0xb9074b0 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7d517d840603fb25 - Init START +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO comm 0xac069d0 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7d517d840603fb25 - Init START +ip-26-0-170-160:65611:66048 [7] NCCL INFO comm 0xa1a1c20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7d517d840603fb25 - Init START +ip-26-0-170-160:65609:66046 [5] NCCL INFO comm 0xab83b60 rank 6 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7d517d840603fb25 - Init START +ip-26-0-170-160:65607:66047 [3] NCCL INFO comm 0xbb7e360 rank 5 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7d517d840603fb25 - Init START +ip-26-0-170-160:65605:66049 [1] NCCL INFO comm 0xbb41230 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7d517d840603fb25 - Init START +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65607:66047 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65605:66049 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65611:66048 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65609:66046 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-170-160:65604:66045 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:65604:66045 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-170-160:65605:66049 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:65605:66049 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-170-160:65606:66043 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:65606:66043 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-170-160:65610:66044 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65610:66044 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-170-160:65607:66047 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:65607:66047 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-170-160:65608:66042 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65608:66042 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-170-160:65609:66046 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65609:66046 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-170-160:65611:66048 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65611:66048 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-170-160:65608:66042 [4] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->2 [5] 7/-1/-1->6->2 [6] -1/-1/-1->6->5 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/2/-1->6->-1 [13] 7/2/-1->6->-1 [14] -1/-1/-1->6->5 [15] -1/-1/-1->6->5 +ip-26-0-170-160:65608:66042 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65610:66044 [6] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 4/-1/-1->7->6 [3] 4/-1/-1->7->6 [4] 4/-1/-1->7->6 [5] 4/-1/-1->7->6 [6] 4/-1/-1->7->3 [7] 4/-1/-1->7->3 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] 4/-1/-1->7->6 [11] 4/-1/-1->7->6 [12] 4/-1/-1->7->6 [13] 4/-1/-1->7->6 [14] 4/3/-1->7->-1 [15] 4/3/-1->7->-1 +ip-26-0-170-160:65610:66044 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO NVLS Head 0: 0 4 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO NVLS Head 1: 1 5 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO NVLS Head 2: 2 6 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO NVLS Head 3: 3 7 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 01/16 : 0 3 2 5 4 7 6 1 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 02/16 : 0 3 6 5 4 7 2 1 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 03/16 : 0 2 7 5 4 6 3 1 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 04/16 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 05/16 : 0 3 2 5 4 7 6 1 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 06/16 : 0 3 6 5 4 7 2 1 +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO NVLS Head 0: 0 4 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 07/16 : 0 2 7 5 4 6 3 1 +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO NVLS Head 1: 1 5 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO NVLS Head 0: 0 4 +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO NVLS Head 2: 2 6 +ip-26-0-170-160:65606:66043 [2] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->1 [3] 6/-1/-1->5->1 [4] -1/-1/-1->5->4 [5] -1/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/1/-1->5->-1 [11] 6/1/-1->5->-1 [12] -1/-1/-1->5->4 [13] -1/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-170-160:65606:66043 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 09/16 : 0 3 2 5 4 7 6 1 +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO NVLS Head 1: 1 5 +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO NVLS Head 3: 3 7 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 10/16 : 0 3 6 5 4 7 2 1 +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO NVLS Head 2: 2 6 +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO NVLS Head 3: 3 7 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 11/16 : 0 2 7 5 4 6 3 1 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 12/16 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 13/16 : 0 3 2 5 4 7 6 1 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 14/16 : 0 3 6 5 4 7 2 1 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 15/16 : 0 2 7 5 4 6 3 1 +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 0/-1/-1->3->2 [3] 0/-1/-1->3->2 [4] 0/-1/-1->3->2 [5] 0/-1/-1->3->2 [6] 0/7/-1->3->-1 [7] 0/7/-1->3->-1 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] 0/-1/-1->3->2 [11] 0/-1/-1->3->2 [12] 0/-1/-1->3->2 [13] 0/-1/-1->3->2 [14] 0/-1/-1->3->7 [15] 0/-1/-1->3->7 +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO NVLS Head 0: 0 4 +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65604:66045 [0] NCCL INFO Trees [0] 5/-1/-1->4->0 [1] 5/-1/-1->4->0 [2] -1/-1/-1->4->7 [3] -1/-1/-1->4->7 [4] 5/-1/-1->4->7 [5] 5/-1/-1->4->7 [6] 5/-1/-1->4->7 [7] 5/-1/-1->4->7 [8] 5/0/-1->4->-1 [9] 5/0/-1->4->-1 [10] -1/-1/-1->4->7 [11] -1/-1/-1->4->7 [12] 5/-1/-1->4->7 [13] 5/-1/-1->4->7 [14] 5/-1/-1->4->7 [15] 5/-1/-1->4->7 +ip-26-0-170-160:65604:66045 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/6/-1->2->-1 [5] 3/6/-1->2->-1 [6] -1/-1/-1->2->1 [7] -1/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->6 [13] 3/-1/-1->2->6 [14] -1/-1/-1->2->1 [15] -1/-1/-1->2->1 +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO NVLS Head 1: 1 5 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Trees [0] 1/4/-1->0->-1 [1] 1/4/-1->0->-1 [2] -1/-1/-1->0->3 [3] -1/-1/-1->0->3 [4] 1/-1/-1->0->3 [5] 1/-1/-1->0->3 [6] 1/-1/-1->0->3 [7] 1/-1/-1->0->3 [8] 1/-1/-1->0->4 [9] 1/-1/-1->0->4 [10] -1/-1/-1->0->3 [11] -1/-1/-1->0->3 [12] 1/-1/-1->0->3 [13] 1/-1/-1->0->3 [14] 1/-1/-1->0->3 [15] 1/-1/-1->0->3 +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO NVLS Head 2: 2 6 +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO NVLS Head 3: 3 7 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/5/-1->1->-1 [3] 2/5/-1->1->-1 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->5 [11] 2/-1/-1->1->5 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 00/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 00/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 04/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 04/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 04/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 08/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 08/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 08/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 12/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 04/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 12/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 12/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 08/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 12/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 00/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 04/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 00/0 : 3[6] -> 4[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 08/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 04/0 : 3[6] -> 4[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 12/0 : 7[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 08/0 : 3[6] -> 4[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 00/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 04/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 00/0 : 7[6] -> 0[0] [send] via NET/Libfabric/0(4)/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 08/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 04/0 : 7[6] -> 0[0] [send] via NET/Libfabric/0(4)/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 12/0 : 3[6] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 00/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 12/0 : 3[6] -> 4[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 08/0 : 7[6] -> 0[0] [send] via NET/Libfabric/0(4)/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 04/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 12/0 : 7[6] -> 0[0] [send] via NET/Libfabric/0(4)/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 08/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 03/0 : 0[0] -> 2[4] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 12/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 07/0 : 0[0] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 11/0 : 0[0] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 15/0 : 0[0] -> 2[4] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 03/0 : 4[0] -> 6[4] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 07/0 : 4[0] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 01/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 02/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 11/0 : 4[0] -> 6[4] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 15/0 : 4[0] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 05/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 06/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 09/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 10/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 13/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 14/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 02/0 : 7[6] -> 2[4] [send] via NET/Libfabric/4(6)/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 06/0 : 7[6] -> 2[4] [send] via NET/Libfabric/4(6)/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 10/0 : 7[6] -> 2[4] [send] via NET/Libfabric/4(6)/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 14/0 : 7[6] -> 2[4] [send] via NET/Libfabric/4(6)/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 01/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 02/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 05/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 06/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 09/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 10/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 13/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 14/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 01/0 : 6[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 05/0 : 6[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 09/0 : 6[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 13/0 : 6[4] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 01/0 : 2[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 05/0 : 2[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 09/0 : 2[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 13/0 : 2[4] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 02/0 : 7[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 06/0 : 7[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 02/0 : 3[6] -> 6[4] [send] via NET/Libfabric/4(2)/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 10/0 : 7[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 06/0 : 3[6] -> 6[4] [send] via NET/Libfabric/4(2)/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 02/0 : 3[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 14/0 : 7[6] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 10/0 : 3[6] -> 6[4] [send] via NET/Libfabric/4(2)/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 06/0 : 3[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 10/0 : 3[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 14/0 : 3[6] -> 6[4] [send] via NET/Libfabric/4(2)/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 14/0 : 3[6] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 01/0 : 2[4] -> 5[2] [send] via NET/Libfabric/2(1)/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 05/0 : 2[4] -> 5[2] [send] via NET/Libfabric/2(1)/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 01/0 : 6[4] -> 1[2] [send] via NET/Libfabric/2(5)/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 09/0 : 2[4] -> 5[2] [send] via NET/Libfabric/2(1)/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 05/0 : 6[4] -> 1[2] [send] via NET/Libfabric/2(5)/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 13/0 : 2[4] -> 5[2] [send] via NET/Libfabric/2(1)/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 09/0 : 6[4] -> 1[2] [send] via NET/Libfabric/2(5)/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 13/0 : 6[4] -> 1[2] [send] via NET/Libfabric/2(5)/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 03/0 : 2[4] -> 7[6] [send] via NET/Libfabric/6(3)/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 03/0 : 6[4] -> 3[6] [send] via NET/Libfabric/6(7)/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 03/0 : 2[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 07/0 : 6[4] -> 3[6] [send] via NET/Libfabric/6(7)/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 07/0 : 2[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 11/0 : 6[4] -> 3[6] [send] via NET/Libfabric/6(7)/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 11/0 : 2[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 15/0 : 6[4] -> 3[6] [send] via NET/Libfabric/6(7)/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 15/0 : 2[4] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 03/0 : 6[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 07/0 : 2[4] -> 7[6] [send] via NET/Libfabric/6(3)/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 07/0 : 6[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 11/0 : 2[4] -> 7[6] [send] via NET/Libfabric/6(3)/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 11/0 : 6[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 15/0 : 2[4] -> 7[6] [send] via NET/Libfabric/6(3)/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 15/0 : 6[4] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 03/0 : 7[6] -> 5[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 07/0 : 7[6] -> 5[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 07/0 : 3[6] -> 1[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 11/0 : 7[6] -> 5[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 11/0 : 3[6] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 15/0 : 3[6] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 15/0 : 7[6] -> 5[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 01/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 02/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 05/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 06/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 09/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 10/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 13/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 14/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 01/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 06/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 05/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 10/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 02/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 09/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 14/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 13/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 03/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 05/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 06/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 07/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 09/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 10/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 11/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 13/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 14/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 15/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->1 [3] 6/-1/-1->5->1 [4] -1/-1/-1->5->4 [5] -1/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/1/-1->5->-1 [11] 6/1/-1->5->-1 [12] -1/-1/-1->5->4 [13] -1/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-170-160:65609:66046 [5] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->2 [5] 7/-1/-1->6->2 [6] -1/-1/-1->6->5 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/2/-1->6->-1 [13] 7/2/-1->6->-1 [14] -1/-1/-1->6->5 [15] -1/-1/-1->6->5 +ip-26-0-170-160:65609:66046 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65607:66047 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO NVLS Head 0: 0 4 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO NVLS Head 1: 1 5 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO NVLS Head 2: 2 6 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO NVLS Head 3: 3 7 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO NVLS Head 0: 0 4 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO NVLS Head 0: 0 4 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/6/-1->2->-1 [5] 3/6/-1->2->-1 [6] -1/-1/-1->2->1 [7] -1/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->6 [13] 3/-1/-1->2->6 [14] -1/-1/-1->2->1 [15] -1/-1/-1->2->1 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO NVLS Head 1: 1 5 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO NVLS Head 1: 1 5 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO NVLS Head 2: 2 6 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO NVLS Head 2: 2 6 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO NVLS Head 3: 3 7 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO NVLS Head 3: 3 7 +ip-26-0-170-160:65611:66048 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 4/-1/-1->7->6 [3] 4/-1/-1->7->6 [4] 4/-1/-1->7->6 [5] 4/-1/-1->7->6 [6] 4/-1/-1->7->3 [7] 4/-1/-1->7->3 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] 4/-1/-1->7->6 [11] 4/-1/-1->7->6 [12] 4/-1/-1->7->6 [13] 4/-1/-1->7->6 [14] 4/3/-1->7->-1 [15] 4/3/-1->7->-1 +ip-26-0-170-160:65611:66048 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO NVLS Head 0: 0 4 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 01/16 : 0 3 2 5 4 7 6 1 +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO NVLS Head 1: 1 5 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 02/16 : 0 3 6 5 4 7 2 1 +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO NVLS Head 2: 2 6 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 03/16 : 0 2 7 5 4 6 3 1 +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO NVLS Head 3: 3 7 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 0/-1/-1->3->2 [3] 0/-1/-1->3->2 [4] 0/-1/-1->3->2 [5] 0/-1/-1->3->2 [6] 0/7/-1->3->-1 [7] 0/7/-1->3->-1 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] 0/-1/-1->3->2 [11] 0/-1/-1->3->2 [12] 0/-1/-1->3->2 [13] 0/-1/-1->3->2 [14] 0/-1/-1->3->7 [15] 0/-1/-1->3->7 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 04/16 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 05/16 : 0 3 2 5 4 7 6 1 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 06/16 : 0 3 6 5 4 7 2 1 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 07/16 : 0 2 7 5 4 6 3 1 +ip-26-0-170-160:65605:66049 [1] NCCL INFO Trees [0] 5/-1/-1->4->0 [1] 5/-1/-1->4->0 [2] -1/-1/-1->4->7 [3] -1/-1/-1->4->7 [4] 5/-1/-1->4->7 [5] 5/-1/-1->4->7 [6] 5/-1/-1->4->7 [7] 5/-1/-1->4->7 [8] 5/0/-1->4->-1 [9] 5/0/-1->4->-1 [10] -1/-1/-1->4->7 [11] -1/-1/-1->4->7 [12] 5/-1/-1->4->7 [13] 5/-1/-1->4->7 [14] 5/-1/-1->4->7 [15] 5/-1/-1->4->7 +ip-26-0-170-160:65605:66049 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 09/16 : 0 3 2 5 4 7 6 1 +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/5/-1->1->-1 [3] 2/5/-1->1->-1 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->5 [11] 2/-1/-1->1->5 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 10/16 : 0 3 6 5 4 7 2 1 +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 11/16 : 0 2 7 5 4 6 3 1 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 12/16 : 0 1 2 3 4 5 6 7 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 13/16 : 0 3 2 5 4 7 6 1 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 14/16 : 0 3 6 5 4 7 2 1 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 15/16 : 0 2 7 5 4 6 3 1 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Trees [0] 1/4/-1->0->-1 [1] 1/4/-1->0->-1 [2] -1/-1/-1->0->3 [3] -1/-1/-1->0->3 [4] 1/-1/-1->0->3 [5] 1/-1/-1->0->3 [6] 1/-1/-1->0->3 [7] 1/-1/-1->0->3 [8] 1/-1/-1->0->4 [9] 1/-1/-1->0->4 [10] -1/-1/-1->0->3 [11] -1/-1/-1->0->3 [12] 1/-1/-1->0->3 [13] 1/-1/-1->0->3 [14] 1/-1/-1->0->3 [15] 1/-1/-1->0->3 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 00/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 04/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 00/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 04/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 04/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 08/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 04/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 08/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 08/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 08/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 12/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 12/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 12/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 12/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 00/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 04/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[1] [send] via NET/Libfabric/1(4)/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 08/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[1] [send] via NET/Libfabric/1(4)/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 12/0 : 3[7] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 00/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[1] [send] via NET/Libfabric/1(4)/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 00/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 04/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 00/0 : 3[7] -> 4[1] [send] via NET/Libfabric/1(0)/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 08/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 04/0 : 3[7] -> 4[1] [send] via NET/Libfabric/1(0)/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 12/0 : 7[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 08/0 : 3[7] -> 4[1] [send] via NET/Libfabric/1(0)/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 12/0 : 3[7] -> 4[1] [send] via NET/Libfabric/1(0)/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 04/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[1] [send] via NET/Libfabric/1(4)/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 08/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 12/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 03/0 : 4[1] -> 6[5] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 07/0 : 4[1] -> 6[5] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 11/0 : 4[1] -> 6[5] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 15/0 : 4[1] -> 6[5] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 01/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 02/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 05/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 06/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 09/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 10/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 13/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 14/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 03/0 : 0[1] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 07/0 : 0[1] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 11/0 : 0[1] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 15/0 : 0[1] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 01/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 02/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 05/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 06/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 09/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 10/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 13/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 14/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 01/0 : 2[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 05/0 : 2[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 09/0 : 2[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 13/0 : 2[5] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 02/0 : 3[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 06/0 : 3[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 02/0 : 7[7] -> 2[5] [send] via NET/Libfabric/5(6)/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 10/0 : 3[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 06/0 : 7[7] -> 2[5] [send] via NET/Libfabric/5(6)/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 14/0 : 3[7] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 10/0 : 7[7] -> 2[5] [send] via NET/Libfabric/5(6)/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 14/0 : 7[7] -> 2[5] [send] via NET/Libfabric/5(6)/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 01/0 : 6[5] -> 1[3] [send] via NET/Libfabric/3(5)/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 05/0 : 6[5] -> 1[3] [send] via NET/Libfabric/3(5)/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 09/0 : 6[5] -> 1[3] [send] via NET/Libfabric/3(5)/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 13/0 : 6[5] -> 1[3] [send] via NET/Libfabric/3(5)/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 01/0 : 6[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 05/0 : 6[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 09/0 : 6[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 13/0 : 6[5] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 02/0 : 7[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 06/0 : 7[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 02/0 : 3[7] -> 6[5] [send] via NET/Libfabric/5(2)/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 10/0 : 7[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 06/0 : 3[7] -> 6[5] [send] via NET/Libfabric/5(2)/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 14/0 : 7[7] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 10/0 : 3[7] -> 6[5] [send] via NET/Libfabric/5(2)/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 14/0 : 3[7] -> 6[5] [send] via NET/Libfabric/5(2)/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 01/0 : 2[5] -> 5[3] [send] via NET/Libfabric/3(1)/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 05/0 : 2[5] -> 5[3] [send] via NET/Libfabric/3(1)/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 09/0 : 2[5] -> 5[3] [send] via NET/Libfabric/3(1)/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 13/0 : 2[5] -> 5[3] [send] via NET/Libfabric/3(1)/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 03/0 : 6[5] -> 3[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Connected all rings +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 01/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 05/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 06/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 07/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 09/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 13/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 14/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 15/0 : 4[0] -> 5[2] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Connected all rings +ip-26-0-170-160:65608:66042 [4] NCCL INFO Connected all rings +ip-26-0-170-160:65610:66044 [6] NCCL INFO Connected all rings +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 01/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 02/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 03/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 05/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 09/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 10/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 11/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 13/0 : 6[4] -> 7[6] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 03/0 : 2[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 07/0 : 6[5] -> 3[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 07/0 : 2[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 11/0 : 6[5] -> 3[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 11/0 : 2[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 15/0 : 6[5] -> 3[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 15/0 : 2[5] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 03/0 : 6[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 03/0 : 2[5] -> 7[7] [send] via NET/Libfabric/7(3)/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 07/0 : 6[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 07/0 : 2[5] -> 7[7] [send] via NET/Libfabric/7(3)/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 11/0 : 6[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 11/0 : 2[5] -> 7[7] [send] via NET/Libfabric/7(3)/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 15/0 : 6[5] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 15/0 : 2[5] -> 7[7] [send] via NET/Libfabric/7(3)/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 03/0 : 7[7] -> 5[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[3] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Connected all rings +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 07/0 : 7[7] -> 5[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 11/0 : 7[7] -> 5[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 15/0 : 7[7] -> 5[3] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 02/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 07/0 : 3[7] -> 1[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 11/0 : 3[7] -> 1[3] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 06/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 10/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 15/0 : 3[7] -> 1[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 14/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 01/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 06/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 10/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 14/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 02/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 01/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 03/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 02/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 06/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 03/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Connected all rings +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 07/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 05/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 09/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 06/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 10/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 07/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 11/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 09/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 14/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 10/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 15/0 : 5[2] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 06/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 11/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 03/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 04/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 02/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 05/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 03/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 12/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 10/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 13/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 11/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 04/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 02/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 05/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 03/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 12/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 10/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 13/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 11/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 07/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 13/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 09/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 04/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 14/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 10/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 07/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 15/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 11/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 11/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 12/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 14/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 15/0 : 1[2] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 15/0 : 4[0] -> 7[6] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 05/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 03/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 06/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 07/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 08/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 14/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 09/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 15/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 06/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 07/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 08/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 14/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 09/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 15/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 09/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 04/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 07/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 10/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 11/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 11/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 13/0 : 2[4] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 12/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 15/0 : 0[0] -> 3[6] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 04/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 05/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 12/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 13/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 04/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 05/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 12/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 02/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 13/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 03/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 10/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 11/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 02/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 03/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 10/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 11/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 06/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 07/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 14/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 15/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 06/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 07/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 14/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 15/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 02/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 08/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 09/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 08/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 09/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 03/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 04/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 00/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 04/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 05/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 04/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 00/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 06/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 08/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 01/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 12/0 : 5[2] -> 4[0] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 07/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 03/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 10/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 07/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 11/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 05/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 08/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 09/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 12/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 11/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 13/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 15/0 : 6[4] -> 5[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 14/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 06/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 15/0 : 7[6] -> 4[0] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 00/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 02/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 07/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 03/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 07/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 04/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 08/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 10/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 08/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 09/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 11/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 11/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 10/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 12/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 15/0 : 2[4] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 13/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 14/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 15/0 : 3[6] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 02/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 03/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 11/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 04/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 08/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 10/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 12/0 : 7[6] -> 6[4] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 11/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 12/0 : 3[6] -> 2[4] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Connected all rings +ip-26-0-170-160:65609:66046 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Connected all rings +ip-26-0-170-160:65605:66049 [1] NCCL INFO Connected all rings +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 01/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 01/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 02/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 03/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 05/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 09/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 10/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 11/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 05/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 06/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 07/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 09/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 13/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 14/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 15/0 : 4[1] -> 5[3] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 13/0 : 6[5] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 06/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 07/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 09/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 01/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 10/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 02/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 11/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 03/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 14/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 06/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 15/0 : 1[3] -> 2[5] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 07/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 09/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 03/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 10/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 11/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 04/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 07/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 14/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 11/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 15/0 : 5[3] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 03/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 12/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 04/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 02/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 05/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 03/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 04/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 12/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 10/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 13/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 05/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 11/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 04/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 02/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 15/0 : 0[1] -> 3[7] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 05/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 03/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 12/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 10/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 13/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 11/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 07/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 11/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 12/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 15/0 : 4[1] -> 7[7] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 09/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 06/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 07/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 08/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 10/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 14/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 09/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 15/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 06/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 08/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 14/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 09/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 11/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-160:65604:66045 [0] NCCL INFO Connected all trees +ip-26-0-170-160:65604:66045 [0] NCCL INFO NVLS comm 0xaf46a60 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 13/0 : 2[5] -> 3[7] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 02/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 04/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 03/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 05/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 10/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 12/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 11/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 13/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 02/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 04/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 03/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO NVLS comm 0xb277560 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 05/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 10/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 12/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 11/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 13/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 00/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 04/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 08/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 06/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 00/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 14/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 08/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 06/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 07/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 14/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 12/0 : 5[3] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 15/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 02/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 01/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO Connected all trees +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 03/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-160:65610:66044 [6] NCCL INFO NVLS comm 0xaba3320 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 09/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 07/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 08/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 09/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 08/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 03/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 09/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 04/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 11/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 05/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 15/0 : 6[5] -> 5[3] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 06/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 07/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-160:65608:66042 [4] NCCL INFO NVLS comm 0xb5724a0 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 10/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 07/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 08/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 09/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 11/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 15/0 : 2[5] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO NVLS comm 0xa9dbbe0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO NVLS comm 0xa3ea720 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[5] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 11/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 12/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 13/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 14/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 15/0 : 7[7] -> 4[1] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[5] via P2P/IPC +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO NVLS comm 0xaf65aa0 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO NVLS comm 0xaf148e0 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 00/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 01/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 00/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 01/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 04/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 05/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 04/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 06/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 05/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 07/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 08/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 08/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 09/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 09/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 10/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 12/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 11/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 13/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 12/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 14/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 13/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 15/0 : 1[2] -> 5[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 00/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 00/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 01/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 01/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 04/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 05/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 04/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 06/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 05/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 07/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 08/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 08/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 09/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 09/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 10/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 12/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 11/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 13/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 12/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 14/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65610:66044 [6] NCCL INFO Channel 13/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65606:66043 [2] NCCL INFO Channel 15/0 : 5[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 00/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 01/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 02/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 03/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 06/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 07/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 08/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 09/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 10/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 11/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 14/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 15/0 : 2[4] -> 6[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 00/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 01/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 02/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 03/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 06/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 07/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 08/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 09/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 10/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 11/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 14/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Channel 15/0 : 6[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 02/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 03/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 04/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 06/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 07/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 10/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 11/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 12/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 14/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 15/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 02/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 03/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 04/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 05/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 06/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 07/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 10/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 11/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 12/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 13/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 14/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66045 [0] NCCL INFO Channel 15/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 00/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 00/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 02/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 01/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 01/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 00/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 03/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 02/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 04/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 01/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 04/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 03/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 02/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 05/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 05/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 04/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 03/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 06/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 06/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 05/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 06/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 07/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 07/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 08/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 07/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 08/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 10/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 09/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 08/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 09/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 11/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 10/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 09/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 12/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 12/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 11/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 10/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 13/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 13/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 12/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 11/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 14/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 14/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 13/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 14/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 15/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 15/0 : 5[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 02/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 15/0 : 6[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 00/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 00/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 03/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 00/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 01/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 01/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 04/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 02/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 04/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 01/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 03/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 05/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 06/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 02/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 04/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 06/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 07/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 03/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 05/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 07/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 10/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 08/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 06/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 08/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 11/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 07/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 09/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 09/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 12/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 10/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 08/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 12/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 11/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 13/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 09/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 14/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 12/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 10/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 14/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Channel 15/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Channel 13/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Channel 15/0 : 1[2] -> 5[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 11/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 14/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Channel 15/0 : 2[4] -> 6[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO NVLS comm 0xbe8cba0 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-160:65607:66047 [3] NCCL INFO Connected all trees +ip-26-0-170-160:65607:66047 [3] NCCL INFO NVLS comm 0xbb7e360 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-160:65605:66049 [1] NCCL INFO Connected all trees +ip-26-0-170-160:65605:66049 [1] NCCL INFO NVLS comm 0xbb41230 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-160:65609:66046 [5] NCCL INFO Connected all trees +ip-26-0-170-160:65609:66046 [5] NCCL INFO NVLS comm 0xab83b60 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-160:65611:66048 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Connected all trees +ip-26-0-170-160:65611:66048 [7] NCCL INFO NVLS comm 0xa1a1c20 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO NVLS comm 0xac069d0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO NVLS comm 0xb5ecab0 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO NVLS comm 0xb9074b0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 02/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 00/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 03/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 00/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 01/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 04/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 01/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 00/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 05/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 02/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 06/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 01/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 03/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 04/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 07/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 04/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 06/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 05/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 10/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 05/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 07/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 08/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 11/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 06/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 08/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 09/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 12/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 07/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 10/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 13/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 09/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 08/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 14/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 09/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 10/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 11/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 12/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 15/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 12/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 11/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 13/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 02/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 13/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 14/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 14/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 03/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 00/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 15/0 : 2[5] -> 6[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 04/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 15/0 : 1[3] -> 5[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 01/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 00/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 05/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 00/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 01/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 06/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 01/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 07/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 02/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 04/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 04/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 10/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 03/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 05/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 11/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 06/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 08/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 12/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 07/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 09/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 13/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 08/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 10/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 14/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 09/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 11/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66049 [1] NCCL INFO Channel 15/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 05/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 10/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 12/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 11/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 06/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65611:66048 [7] NCCL INFO Channel 13/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 14/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 07/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65609:66046 [5] NCCL INFO Channel 15/0 : 6[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 08/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 09/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 12/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 13/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 14/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-160:65607:66047 [3] NCCL INFO Channel 15/0 : 5[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 00/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 00/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 01/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 01/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 02/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 04/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 00/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 02/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 03/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 05/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 01/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 03/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 04/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 02/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 06/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 04/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 05/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 03/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 05/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 06/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 07/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 06/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 07/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 08/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 08/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 07/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 10/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 09/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 09/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 08/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 11/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 12/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 10/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 09/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 12/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 13/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 11/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 10/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 13/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 12/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 14/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 11/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 14/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 15/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 15/0 : 5[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 13/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 14/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 02/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 00/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 00/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 15/0 : 6[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 03/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 01/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 01/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 00/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 04/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 04/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 01/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 02/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 05/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 05/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 03/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 06/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 02/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 06/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 04/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 03/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 07/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 07/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 05/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 06/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 08/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 10/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 08/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 07/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 09/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 11/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 09/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 08/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 12/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 10/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 12/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 09/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 13/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 11/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 10/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 13/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 14/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 12/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 11/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 14/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Channel 15/0 : 1[3] -> 5[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Channel 13/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Channel 15/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 14/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Channel 15/0 : 2[5] -> 6[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-170-160:65608:66042 [4] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65608:66042 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65608:66042 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65608:66042 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65606:66043 [2] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65606:66043 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65606:66043 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65606:66043 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65610:66044 [6] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65610:66044 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65610:66044 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65610:66044 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65604:66045 [0] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65604:66045 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65604:66045 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65604:66045 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65610:66044 [6] NCCL INFO comm 0xaba3320 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb259cfe95f7000d5 - Init COMPLETE +ip-26-0-170-160:65606:66043 [2] NCCL INFO comm 0xb277560 rank 5 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb259cfe95f7000d5 - Init COMPLETE +ip-26-0-170-160:65604:66045 [0] NCCL INFO comm 0xaf46a60 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb259cfe95f7000d5 - Init COMPLETE +ip-26-0-170-160:65608:66042 [4] NCCL INFO comm 0xb5724a0 rank 6 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb259cfe95f7000d5 - Init COMPLETE +ip-26-0-170-143:2167363:2167804 [0] NCCL INFO comm 0xa9dbbe0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb259cfe95f7000d5 - Init COMPLETE +ip-26-0-170-143:2167367:2167806 [4] NCCL INFO comm 0xaf148e0 rank 2 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb259cfe95f7000d5 - Init COMPLETE +ip-26-0-170-143:2167365:2167807 [2] NCCL INFO comm 0xaf65aa0 rank 1 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb259cfe95f7000d5 - Init COMPLETE +ip-26-0-170-143:2167369:2167805 [6] NCCL INFO comm 0xa3ea720 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb259cfe95f7000d5 - Init COMPLETE +12/28/2024 00:53:38 [INFO|DP=0|PP=3|TP=0|ip-26-0-170-143]: Local number of parameters: 453M (864.05MiB) +12/28/2024 00:53:38 [INFO|DP=0|PP=2|TP=0|ip-26-0-170-143]: Local number of parameters: 453M (864.05MiB) +12/28/2024 00:53:38 [INFO|DP=0|PP=6|TP=0|ip-26-0-170-160]: Local number of parameters: 453M (864.05MiB) +12/28/2024 00:53:38 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: Local number of parameters: 403M (768.01MiB) +12/28/2024 00:53:38 [INFO|DP=0|PP=4|TP=0|ip-26-0-170-160]: Local number of parameters: 453M (864.05MiB) +12/28/2024 00:53:38 [INFO|DP=0|PP=5|TP=0|ip-26-0-170-160]: Local number of parameters: 453M (864.05MiB) +12/28/2024 00:53:38 [INFO|DP=0|PP=3|TP=0|ip-26-0-170-143]: [After model building] Memory usage: 864.06MiB. Peak allocated: 5504.00MiB Peak reserved: 12674.00MiB +12/28/2024 00:53:38 [INFO|DP=0|PP=6|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 864.06MiB. Peak allocated: 5504.00MiB Peak reserved: 12674.00MiB +12/28/2024 00:53:38 [INFO|DP=0|PP=2|TP=0|ip-26-0-170-143]: [After model building] Memory usage: 864.06MiB. Peak allocated: 5504.00MiB Peak reserved: 12674.00MiB +12/28/2024 00:53:38 [INFO|DP=0|PP=7|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 768.02MiB. Peak allocated: 5504.00MiB Peak reserved: 12674.00MiB +12/28/2024 00:53:38 [INFO|DP=0|PP=5|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 864.06MiB. Peak allocated: 5504.00MiB Peak reserved: 12674.00MiB +12/28/2024 00:53:38 [INFO|DP=0|PP=4|TP=0|ip-26-0-170-160]: [After model building] Memory usage: 864.06MiB. Peak allocated: 5504.00MiB Peak reserved: 12674.00MiB +12/28/2024 00:53:38 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-143]: Local number of parameters: 453M (864.05MiB) +12/28/2024 00:53:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Total number of parameters: 3.98G (7584.33MiB) +12/28/2024 00:53:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Local number of parameters: 856M (1632.05MiB) +12/28/2024 00:53:38 [INFO|DP=0|PP=1|TP=0|ip-26-0-170-143]: [After model building] Memory usage: 864.06MiB. Peak allocated: 5504.00MiB Peak reserved: 12674.00MiB +12/28/2024 00:53:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [After model building] Memory usage: 1632.06MiB. Peak allocated: 5504.00MiB Peak reserved: 12674.00MiB +12/28/2024 00:53:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: No checkpoint path provided. +12/28/2024 00:53:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Parametrizing model parameters using StandardParametrizator +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65611:66048 [7] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65611:66048 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65611:66048 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65611:66048 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65605:66049 [1] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65605:66049 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65605:66049 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65605:66049 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65607:66047 [3] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65607:66047 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65607:66047 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65607:66047 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65609:66046 [5] NCCL INFO Connected NVLS tree +ip-26-0-170-160:65609:66046 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65609:66046 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-160:65609:66046 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65605:66049 [1] NCCL INFO comm 0xbb41230 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7d517d840603fb25 - Init COMPLETE +ip-26-0-170-160:65609:66046 [5] NCCL INFO comm 0xab83b60 rank 6 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7d517d840603fb25 - Init COMPLETE +ip-26-0-170-160:65611:66048 [7] NCCL INFO comm 0xa1a1c20 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7d517d840603fb25 - Init COMPLETE +ip-26-0-170-160:65607:66047 [3] NCCL INFO comm 0xbb7e360 rank 5 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7d517d840603fb25 - Init COMPLETE +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO Connected NVLS tree +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167366:2167810 [3] NCCL INFO comm 0xbe8cba0 rank 1 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7d517d840603fb25 - Init COMPLETE +ip-26-0-170-143:2167370:2167812 [7] NCCL INFO comm 0xb9074b0 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7d517d840603fb25 - Init COMPLETE +ip-26-0-170-143:2167364:2167809 [1] NCCL INFO comm 0xac069d0 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7d517d840603fb25 - Init COMPLETE +ip-26-0-170-143:2167368:2167811 [5] NCCL INFO comm 0xb5ecab0 rank 2 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7d517d840603fb25 - Init COMPLETE +ip-26-0-170-160:65610:66077 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:65611:66078 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:65611:66078 [7] NCCL INFO comm 0xa36c530 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4da6a88953e4b81e - Init START +ip-26-0-170-160:65610:66077 [6] NCCL INFO comm 0xad6e180 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4da6a88953e4b81e - Init START +ip-26-0-170-160:65611:66078 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65610:66077 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:65609:66081 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:65606:66085 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:65604:66084 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:65605:66086 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:65609:66081 [5] NCCL INFO comm 0xad50e10 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7599c7f07325373e - Init START +ip-26-0-170-160:65608:66080 [4] NCCL INFO comm 0xb73cc90 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7599c7f07325373e - Init START +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:65609:66081 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:65608:66080 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO comm 0xa9ee8a0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x79e350c6ca75ea22 - Init START +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO comm 0xac19690 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x79e350c6ca75ea22 - Init START +ip-26-0-170-160:65607:66087 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-160:65604:66084 [0] NCCL INFO comm 0xb1113e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x74b53129d4011cee - Init START +ip-26-0-170-160:65605:66086 [1] NCCL INFO comm 0xbd0aa20 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x74b53129d4011cee - Init START +ip-26-0-170-160:65604:66084 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:65605:66086 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO comm 0xb1306d0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1b32842ae1902bc - Init START +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO comm 0xc057f50 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1b32842ae1902bc - Init START +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO comm 0xb0e1240 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd7ad218274afa395 - Init START +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO comm 0xb7b6970 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd7ad218274afa395 - Init START +ip-26-0-170-160:65607:66087 [3] NCCL INFO comm 0xbd49490 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5a474317311fb7fa - Init START +ip-26-0-170-160:65606:66085 [2] NCCL INFO comm 0xb443430 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5a474317311fb7fa - Init START +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65607:66087 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65606:66085 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO comm 0xbad21a0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3dbb1c9c58363080 - Init START +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO comm 0xa5b60f0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x3dbb1c9c58363080 - Init START +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65611:66078 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-160:65611:66078 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-160:65611:66078 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-160:65610:66077 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:65609:66081 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:65605:66086 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-160:65605:66086 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-160:65605:66086 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-160:65604:66084 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-160:65609:66081 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-160:65609:66081 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-160:65608:66080 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65607:66087 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-160:65607:66087 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-160:65607:66087 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-160:65606:66085 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65610:66077 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65605:66086 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65604:66084 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66081 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66087 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 04/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 05/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 06/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 07/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 08/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 09/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 10/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 11/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 12/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 13/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 14/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 15/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 16/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65606:66085 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 17/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 18/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 19/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 20/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 21/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 22/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Channel 23/0 : 0[6] -> 1[7] via P2P/IPC +ip-26-0-170-160:65608:66080 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[6] via P2P/IPC +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-170-160:65611:66078 [7] NCCL INFO Connected all rings +ip-26-0-170-160:65611:66078 [7] NCCL INFO Connected all trees +ip-26-0-170-160:65611:66078 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65611:66078 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65611:66078 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65610:66077 [6] NCCL INFO Connected all rings +ip-26-0-170-160:65610:66077 [6] NCCL INFO Connected all trees +ip-26-0-170-160:65610:66077 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65610:66077 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65610:66077 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65611:66078 [7] NCCL INFO comm 0xa36c530 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4da6a88953e4b81e - Init COMPLETE +ip-26-0-170-160:65610:66077 [6] NCCL INFO comm 0xad6e180 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4da6a88953e4b81e - Init COMPLETE +ip-26-0-170-160:65605:66086 [1] NCCL INFO Connected all rings +ip-26-0-170-160:65605:66086 [1] NCCL INFO Connected all trees +ip-26-0-170-160:65605:66086 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65605:66086 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65605:66086 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65604:66084 [0] NCCL INFO Connected all rings +ip-26-0-170-160:65604:66084 [0] NCCL INFO Connected all trees +ip-26-0-170-160:65604:66084 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65604:66084 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65604:66084 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65604:66084 [0] NCCL INFO comm 0xb1113e0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x74b53129d4011cee - Init COMPLETE +ip-26-0-170-160:65605:66086 [1] NCCL INFO comm 0xbd0aa20 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x74b53129d4011cee - Init COMPLETE +ip-26-0-170-160:65607:66087 [3] NCCL INFO Connected all rings +ip-26-0-170-160:65607:66087 [3] NCCL INFO Connected all trees +ip-26-0-170-160:65607:66087 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65607:66087 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65607:66087 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65608:66080 [4] NCCL INFO Connected all rings +ip-26-0-170-160:65608:66080 [4] NCCL INFO Connected all trees +ip-26-0-170-160:65608:66080 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65608:66080 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65608:66080 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65606:66085 [2] NCCL INFO Connected all rings +ip-26-0-170-160:65606:66085 [2] NCCL INFO Connected all trees +ip-26-0-170-160:65606:66085 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65606:66085 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65606:66085 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65609:66081 [5] NCCL INFO Connected all rings +ip-26-0-170-160:65609:66081 [5] NCCL INFO Connected all trees +ip-26-0-170-160:65609:66081 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65609:66081 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65609:66081 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65607:66087 [3] NCCL INFO comm 0xbd49490 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5a474317311fb7fa - Init COMPLETE +ip-26-0-170-160:65606:66085 [2] NCCL INFO comm 0xb443430 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5a474317311fb7fa - Init COMPLETE +ip-26-0-170-160:65608:66080 [4] NCCL INFO comm 0xb73cc90 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7599c7f07325373e - Init COMPLETE +ip-26-0-170-160:65609:66081 [5] NCCL INFO comm 0xad50e10 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7599c7f07325373e - Init COMPLETE +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167370:2167851 [7] NCCL INFO comm 0xbad21a0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3dbb1c9c58363080 - Init COMPLETE +ip-26-0-170-143:2167369:2167849 [6] NCCL INFO comm 0xa5b60f0 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x3dbb1c9c58363080 - Init COMPLETE +ip-26-0-170-143:2167365:2167845 [2] NCCL INFO comm 0xb1306d0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1b32842ae1902bc - Init COMPLETE +ip-26-0-170-143:2167366:2167848 [3] NCCL INFO comm 0xc057f50 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1b32842ae1902bc - Init COMPLETE +ip-26-0-170-143:2167368:2167850 [5] NCCL INFO comm 0xb7b6970 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd7ad218274afa395 - Init COMPLETE +ip-26-0-170-143:2167367:2167846 [4] NCCL INFO comm 0xb0e1240 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd7ad218274afa395 - Init COMPLETE +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167363:2167841 [0] NCCL INFO comm 0xa9ee8a0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x79e350c6ca75ea22 - Init COMPLETE +ip-26-0-170-143:2167364:2167842 [1] NCCL INFO comm 0xac19690 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x79e350c6ca75ea22 - Init COMPLETE +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:65610:66107 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:65611:66108 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO comm 0xaa19a80 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7e2be6c9d9ca9af0 - Init START +ip-26-0-170-160:65610:66107 [6] NCCL INFO comm 0xad920d0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7e2be6c9d9ca9af0 - Init START +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO comm 0xac44790 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9b464eb49dafa053 - Init START +ip-26-0-170-160:65611:66108 [7] NCCL INFO comm 0xa3901b0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9b464eb49dafa053 - Init START +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65610:66107 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65611:66108 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:65610:66107 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65611:66108 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65610:66107 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:65610:66107 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65611:66108 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:65611:66108 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-160:65610:66107 [6] NCCL INFO Channel 00/0 : 0[0] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65610:66107 [6] NCCL INFO Channel 01/0 : 0[0] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 00/0 : 1[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:66107 [6] NCCL INFO Channel 02/0 : 0[0] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 01/0 : 1[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:66107 [6] NCCL INFO Channel 03/0 : 0[0] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 02/0 : 1[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:66107 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 03/0 : 1[7] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:66107 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:66107 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65610:66107 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 00/0 : 1[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 01/0 : 1[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 02/0 : 1[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 03/0 : 1[6] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65611:66108 [7] NCCL INFO Channel 00/0 : 0[1] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:66108 [7] NCCL INFO Channel 01/0 : 0[1] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:66108 [7] NCCL INFO Channel 02/0 : 0[1] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:66108 [7] NCCL INFO Channel 03/0 : 0[1] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:66108 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:66108 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:66108 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65611:66108 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[1] [send] via NET/Libfabric/7/GDRDMA +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167363:2167873 [0] NCCL INFO comm 0xaa19a80 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7e2be6c9d9ca9af0 - Init COMPLETE +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Optimizer Building] Using LearningRateForSP as learning rate +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] Size of optimizer params per rank: +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] DP Rank 0 has 428M out of 856M (50.00%) params' optimizer states +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [ZeRO sharding] DP Rank 1 has 428M out of 856M (50.00%) params' optimizer states +ip-26-0-170-143:2167364:2167874 [1] NCCL INFO comm 0xac44790 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9b464eb49dafa053 - Init COMPLETE +ip-26-0-170-160:65611:66108 [7] NCCL INFO Connected all rings +ip-26-0-170-160:65611:66108 [7] NCCL INFO Connected all trees +ip-26-0-170-160:65611:66108 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65611:66108 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65611:66108 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65610:66107 [6] NCCL INFO Connected all rings +ip-26-0-170-160:65610:66107 [6] NCCL INFO Connected all trees +ip-26-0-170-160:65610:66107 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65610:66107 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65610:66107 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65611:66108 [7] NCCL INFO comm 0xa3901b0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9b464eb49dafa053 - Init COMPLETE +ip-26-0-170-160:65610:66107 [6] NCCL INFO comm 0xad920d0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7e2be6c9d9ca9af0 - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.token_position_embeddings | PP: 0/8 | Block rank: 0 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.0 | PP: 0/8 | Block rank: 1 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.1 | PP: 0/8 | Block rank: 2 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.2 | PP: 0/8 | Block rank: 3 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.3 | PP: 0/8 | Block rank: 4 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.4 | PP: 1/8 | Block rank: 0 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.5 | PP: 1/8 | Block rank: 1 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.6 | PP: 1/8 | Block rank: 2 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.7 | PP: 1/8 | Block rank: 3 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.8 | PP: 2/8 | Block rank: 0 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.9 | PP: 2/8 | Block rank: 1 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.10 | PP: 2/8 | Block rank: 2 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.11 | PP: 2/8 | Block rank: 3 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.12 | PP: 3/8 | Block rank: 0 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.13 | PP: 3/8 | Block rank: 1 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.14 | PP: 3/8 | Block rank: 2 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.15 | PP: 3/8 | Block rank: 3 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.16 | PP: 4/8 | Block rank: 0 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.17 | PP: 4/8 | Block rank: 1 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.18 | PP: 4/8 | Block rank: 2 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.19 | PP: 4/8 | Block rank: 3 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.20 | PP: 5/8 | Block rank: 0 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.21 | PP: 5/8 | Block rank: 1 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.22 | PP: 5/8 | Block rank: 2 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.23 | PP: 5/8 | Block rank: 3 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.24 | PP: 6/8 | Block rank: 0 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.25 | PP: 6/8 | Block rank: 1 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.26 | PP: 6/8 | Block rank: 2 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.decoder.27 | PP: 6/8 | Block rank: 3 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.final_layer_norm | PP: 7/8 | Block rank: 0 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.lm_head | PP: 7/8 | Block rank: 1 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: model.cast_to_fp32 | PP: 7/8 | Block rank: 2 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: module_name: loss | PP: 7/8 | Block rank: 3 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Using dummy data generator +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Training Plan] There are 1 training stages +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Stage Stable Training Stage] start from step 1 +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: [Start training] datetime: 2024-12-28 00:53:41.452749 | mbs: 4 | grad_accum: 32 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/28/2024 00:53:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-170-143]: Memory usage: 6528.20MiB. Peak allocated 6528.20MiB. Peak reserved: 17574.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO comm 0x17b535a0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa40db374750400e2 - Init START +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO comm 0xb3a6f00 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa40db374750400e2 - Init START +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO comm 0x184b1ea0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x21143772262faeee - Init START +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO comm 0xc2df5f0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x21143772262faeee - Init START +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Connected all rings +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO Connected all trees +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167363:2167887 [0] NCCL INFO comm 0x17b535a0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa40db374750400e2 - Init COMPLETE +ip-26-0-170-143:2167365:2167888 [2] NCCL INFO comm 0xb3a6f00 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa40db374750400e2 - Init COMPLETE +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 00/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 01/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 04/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 05/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 06/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 07/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Connected all rings +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO Connected all trees +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 08/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 09/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 10/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 11/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 12/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 13/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 14/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 15/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 16/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 17/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 18/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 19/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 20/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 21/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 22/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 23/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 24/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 25/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 26/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 27/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 28/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 29/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 30/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167363:2167899 [0] NCCL INFO Channel 31/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-143:2167366:2167891 [3] NCCL INFO comm 0xc2df5f0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x21143772262faeee - Init COMPLETE +ip-26-0-170-143:2167364:2167890 [1] NCCL INFO comm 0x184b1ea0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x21143772262faeee - Init COMPLETE +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 00/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 01/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 04/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 05/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 06/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 07/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 08/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 09/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 10/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 11/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 12/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 13/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 14/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 15/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 16/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 17/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 18/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 19/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 20/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 21/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 22/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 23/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 24/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 25/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 26/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 27/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 28/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 29/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 30/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-143:2167364:2167905 [1] NCCL INFO Channel 31/1 : 0[1] -> 1[3] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO comm 0x13c408a0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb1d8a5c2f31ed04e - Init START +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO comm 0xb352f30 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb1d8a5c2f31ed04e - Init START +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 04/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 05/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 06/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 07/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 08/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 09/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 10/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 11/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 12/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 13/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 14/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 15/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 16/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 17/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 18/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 19/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 20/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 21/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 22/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Channel 23/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Connected all rings +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO Connected all trees +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167367:2167913 [4] NCCL INFO comm 0xb352f30 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb1d8a5c2f31ed04e - Init COMPLETE +ip-26-0-170-143:2167365:2167912 [2] NCCL INFO comm 0x13c408a0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb1d8a5c2f31ed04e - Init COMPLETE +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 00/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 01/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 04/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 05/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 06/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 07/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 08/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 09/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 10/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 11/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 12/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 13/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 14/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 15/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 16/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 17/1 : 0[2] -> 1[4] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 18/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 19/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 20/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 21/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO comm 0xba61550 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8adccfaf14e4d6f1 - Init START +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO comm 0x148b5420 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8adccfaf14e4d6f1 - Init START +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 22/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 23/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 24/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 25/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 26/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 27/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 28/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 29/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 30/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167365:2167920 [2] NCCL INFO Channel 31/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 04/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 05/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 06/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 07/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 08/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 09/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 10/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 11/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 12/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 13/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 14/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 15/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 16/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 17/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 18/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 19/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 20/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 21/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 22/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Channel 23/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Connected all rings +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO Connected all trees +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167366:2167922 [3] NCCL INFO comm 0x148b5420 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8adccfaf14e4d6f1 - Init COMPLETE +ip-26-0-170-143:2167368:2167923 [5] NCCL INFO comm 0xba61550 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8adccfaf14e4d6f1 - Init COMPLETE +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 00/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 01/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 04/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 05/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 06/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 07/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 08/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 09/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 10/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 11/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 12/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 13/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 14/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 15/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 16/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 17/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 18/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 19/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 20/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 21/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 22/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 23/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 24/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 25/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 26/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 27/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 28/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 29/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 30/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-143:2167366:2167930 [3] NCCL INFO Channel 31/1 : 0[3] -> 1[5] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO comm 0xa862370 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x81101b7da0c80506 - Init START +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO comm 0x1394fde0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x81101b7da0c80506 - Init START +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 04/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 05/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 06/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 07/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 08/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 09/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 10/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 11/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 12/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 13/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 14/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 15/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 16/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 17/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 18/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 19/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 20/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 21/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 22/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Channel 23/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Connected all rings +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO Connected all trees +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167369:2167939 [6] NCCL INFO comm 0xa862370 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x81101b7da0c80506 - Init COMPLETE +ip-26-0-170-143:2167367:2167938 [4] NCCL INFO comm 0x1394fde0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x81101b7da0c80506 - Init COMPLETE +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 00/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 01/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 04/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 05/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 06/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 07/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 08/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 09/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 10/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 11/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 12/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 13/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 14/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 15/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 16/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 17/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 18/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 19/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 20/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 21/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 22/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 23/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 24/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 25/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 26/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 27/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 28/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 29/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 30/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-143:2167367:2167945 [4] NCCL INFO Channel 31/1 : 0[4] -> 1[6] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO comm 0x14ca2ea0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6e12d9a8ca8ad56f - Init START +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO comm 0xbd441d0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6e12d9a8ca8ad56f - Init START +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 04/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 05/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 06/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 07/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 08/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 09/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 10/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 11/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 12/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 13/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 14/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 15/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 16/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 17/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 18/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 19/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 20/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 21/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 22/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Channel 23/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Connected all rings +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO Connected all trees +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-143:2167370:2167949 [7] NCCL INFO comm 0xbd441d0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6e12d9a8ca8ad56f - Init COMPLETE +ip-26-0-170-143:2167368:2167948 [5] NCCL INFO comm 0x14ca2ea0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6e12d9a8ca8ad56f - Init COMPLETE +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 00/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 01/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 04/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 05/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 06/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 07/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 08/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 09/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 10/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 11/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 12/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 13/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 14/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 15/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 16/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 17/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 18/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 19/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 20/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 21/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 22/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 23/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 24/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 25/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 26/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 27/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 28/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 29/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 30/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-143:2167368:2167954 [5] NCCL INFO Channel 31/1 : 0[5] -> 1[7] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:65604:66136 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:65604:66136 [0] NCCL INFO comm 0xb388930 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5f1f6347ae759e36 - Init START +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO comm 0x14b43260 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5f1f6347ae759e36 - Init START +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65604:66136 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65604:66136 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:65604:66136 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:65604:66136 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 00/0 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 01/0 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 02/0 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 03/0 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-170-160:65604:66136 [0] NCCL INFO Channel 00/0 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66136 [0] NCCL INFO Channel 01/0 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66136 [0] NCCL INFO Channel 02/0 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66136 [0] NCCL INFO Channel 03/0 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66136 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66136 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66136 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66136 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-170-160:65604:66136 [0] NCCL INFO Connected all rings +ip-26-0-170-160:65604:66136 [0] NCCL INFO Connected all trees +ip-26-0-170-160:65604:66136 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65604:66136 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65604:66136 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Connected all rings +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO Connected all trees +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65604:66136 [0] NCCL INFO comm 0xb388930 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5f1f6347ae759e36 - Init COMPLETE +ip-26-0-170-160:65604:66140 [0] NCCL INFO Channel 02/1 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-160:65604:66140 [0] NCCL INFO Channel 03/1 : 0[6] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-143:2167369:2167963 [6] NCCL INFO comm 0x14b43260 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5f1f6347ae759e36 - Init COMPLETE +ip-26-0-170-143:2167369:2167967 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2167369:2167967 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[0] [send] via NET/Libfabric/6/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:65605:66141 [1] NCCL INFO Using network Libfabric +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO comm 0x1420a5b0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7eeb41d9e718b5a9 - Init START +ip-26-0-170-160:65605:66141 [1] NCCL INFO comm 0xbfdb7c0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7eeb41d9e718b5a9 - Init START +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65605:66141 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65605:66141 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:65605:66141 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-170-160:65605:66141 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 00/0 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 01/0 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 02/0 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 03/0 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-170-160:65605:66141 [1] NCCL INFO Channel 00/0 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:66141 [1] NCCL INFO Channel 01/0 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:66141 [1] NCCL INFO Channel 02/0 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:66141 [1] NCCL INFO Channel 03/0 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:66141 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:66141 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:66141 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-160:65605:66141 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Connected all rings +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO Connected all trees +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-160:65605:66141 [1] NCCL INFO Connected all rings +ip-26-0-170-160:65605:66141 [1] NCCL INFO Connected all trees +ip-26-0-170-160:65605:66141 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65605:66141 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65605:66141 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-170-143:2167370:2167969 [7] NCCL INFO comm 0x1420a5b0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7eeb41d9e718b5a9 - Init COMPLETE +ip-26-0-170-143:2167370:2167972 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2167370:2167972 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[1] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-160:65605:66141 [1] NCCL INFO comm 0xbfdb7c0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7eeb41d9e718b5a9 - Init COMPLETE +ip-26-0-170-160:65605:66146 [1] NCCL INFO Channel 02/1 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-160:65605:66146 [1] NCCL INFO Channel 03/1 : 0[7] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:65604:66152 [0] NCCL INFO Using network Libfabric +ip-26-0-170-160:65606:66153 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:65606:66153 [2] NCCL INFO comm 0xb6bd330 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xaf908b3bc1d616a8 - Init START +ip-26-0-170-160:65604:66152 [0] NCCL INFO comm 0x155667a0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xaf908b3bc1d616a8 - Init START +ip-26-0-170-160:65606:66153 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65604:66152 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65606:66153 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:65604:66152 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-160:65606:66153 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-160:65606:66153 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-160:65604:66152 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 04/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 05/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 06/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 07/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 08/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 09/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 10/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 11/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 12/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 13/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 14/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 15/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 16/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 17/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 18/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 19/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 20/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 21/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 22/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Channel 23/0 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65604:66152 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65606:66153 [2] NCCL INFO Connected all rings +ip-26-0-170-160:65606:66153 [2] NCCL INFO Connected all trees +ip-26-0-170-160:65606:66153 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65606:66153 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65606:66153 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65604:66152 [0] NCCL INFO Connected all rings +ip-26-0-170-160:65604:66152 [0] NCCL INFO Connected all trees +ip-26-0-170-160:65604:66152 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65604:66152 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65604:66152 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65606:66153 [2] NCCL INFO comm 0xb6bd330 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xaf908b3bc1d616a8 - Init COMPLETE +ip-26-0-170-160:65604:66152 [0] NCCL INFO comm 0x155667a0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xaf908b3bc1d616a8 - Init COMPLETE +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 00/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 01/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 04/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 05/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 06/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 07/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 08/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 09/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 10/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 11/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 12/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 13/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 14/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 15/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 16/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 17/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 18/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 19/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 20/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 21/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 22/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 23/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 24/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 25/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 26/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 27/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 28/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 29/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 30/1 : 0[0] -> 1[2] via P2P/IPC +ip-26-0-170-160:65604:66159 [0] NCCL INFO Channel 31/1 : 0[0] -> 1[2] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:65605:66162 [1] NCCL INFO Using network Libfabric +ip-26-0-170-160:65607:66163 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:65605:66162 [1] NCCL INFO comm 0x147adef0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8d5df7a7b06205d0 - Init START +ip-26-0-170-160:65607:66163 [3] NCCL INFO comm 0xbfc34e0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8d5df7a7b06205d0 - Init START +ip-26-0-170-160:65607:66163 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65605:66162 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65607:66163 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:65605:66162 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-160:65607:66163 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-160:65607:66163 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65605:66162 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-160:65605:66162 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 04/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 05/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 06/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 07/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 08/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 09/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 10/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 11/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 12/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 13/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 14/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 15/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 16/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 17/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 18/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 19/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 20/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 21/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 22/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Channel 23/0 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65605:66162 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65607:66163 [3] NCCL INFO Connected all rings +ip-26-0-170-160:65607:66163 [3] NCCL INFO Connected all trees +ip-26-0-170-160:65607:66163 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65607:66163 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65607:66163 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65605:66162 [1] NCCL INFO Connected all rings +ip-26-0-170-160:65605:66162 [1] NCCL INFO Connected all trees +ip-26-0-170-160:65605:66162 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65605:66162 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65605:66162 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65607:66163 [3] NCCL INFO comm 0xbfc34e0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x8d5df7a7b06205d0 - Init COMPLETE +ip-26-0-170-160:65605:66162 [1] NCCL INFO comm 0x147adef0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x8d5df7a7b06205d0 - Init COMPLETE +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 00/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 01/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 04/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 05/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 06/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 07/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 08/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 09/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 10/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 11/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 12/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 13/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 14/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 15/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 16/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 17/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 18/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 19/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 20/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 21/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 22/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 23/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 24/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 25/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 26/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 27/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 28/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 29/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 30/1 : 0[1] -> 1[3] via P2P/IPC +ip-26-0-170-160:65605:66169 [1] NCCL INFO Channel 31/1 : 0[1] -> 1[3] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:65606:66178 [2] NCCL INFO Using network Libfabric +ip-26-0-170-160:65608:66179 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:65608:66179 [4] NCCL INFO comm 0xb9bd790 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb52b2dc4d22ae9a8 - Init START +ip-26-0-170-160:65606:66178 [2] NCCL INFO comm 0x13cb1c90 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb52b2dc4d22ae9a8 - Init START +ip-26-0-170-160:65608:66179 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65606:66178 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-170-160:65608:66179 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-160:65608:66179 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-160:65608:66179 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-160:65606:66178 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 04/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 05/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 06/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 07/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 08/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 09/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 10/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 11/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 12/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 13/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 14/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 15/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 16/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 17/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 18/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 19/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 20/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 21/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 22/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65606:66178 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Channel 23/0 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66179 [4] NCCL INFO Connected all rings +ip-26-0-170-160:65608:66179 [4] NCCL INFO Connected all trees +ip-26-0-170-160:65608:66179 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65608:66179 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65608:66179 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65606:66178 [2] NCCL INFO Connected all rings +ip-26-0-170-160:65606:66178 [2] NCCL INFO Connected all trees +ip-26-0-170-160:65606:66178 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65606:66178 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65606:66178 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65608:66179 [4] NCCL INFO comm 0xb9bd790 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb52b2dc4d22ae9a8 - Init COMPLETE +ip-26-0-170-160:65606:66178 [2] NCCL INFO comm 0x13cb1c90 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb52b2dc4d22ae9a8 - Init COMPLETE +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 00/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 01/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 04/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 05/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 06/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 07/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 08/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 09/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 10/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 11/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 12/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 13/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 14/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 15/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 16/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 17/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 18/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 19/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 20/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 21/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 22/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 23/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 24/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 25/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 26/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 27/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 28/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 29/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 30/1 : 0[2] -> 1[4] via P2P/IPC +ip-26-0-170-160:65606:66184 [2] NCCL INFO Channel 31/1 : 0[2] -> 1[4] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:65609:66188 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:65607:66187 [3] NCCL INFO Using network Libfabric +ip-26-0-170-160:65607:66187 [3] NCCL INFO comm 0x15dfc2d0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc25286a5b79e2b21 - Init START +ip-26-0-170-160:65609:66188 [5] NCCL INFO comm 0xafc68e0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc25286a5b79e2b21 - Init START +ip-26-0-170-160:65609:66188 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65607:66187 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65609:66188 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-170-160:65609:66188 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-160:65609:66188 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-160:65607:66187 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 04/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 05/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 06/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 07/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 08/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 09/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 10/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 11/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 12/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 13/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 14/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 15/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 16/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 17/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 18/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 19/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 20/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 21/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 22/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66187 [3] NCCL INFO Channel 23/0 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65609:66188 [5] NCCL INFO Connected all rings +ip-26-0-170-160:65609:66188 [5] NCCL INFO Connected all trees +ip-26-0-170-160:65609:66188 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65609:66188 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65609:66188 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65607:66187 [3] NCCL INFO Connected all rings +ip-26-0-170-160:65607:66187 [3] NCCL INFO Connected all trees +ip-26-0-170-160:65607:66187 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65607:66187 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65607:66187 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65609:66188 [5] NCCL INFO comm 0xafc68e0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc25286a5b79e2b21 - Init COMPLETE +ip-26-0-170-160:65607:66187 [3] NCCL INFO comm 0x15dfc2d0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc25286a5b79e2b21 - Init COMPLETE +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 00/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 01/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 04/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 05/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 06/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 07/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 08/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 09/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 10/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 11/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 12/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 13/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 14/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 15/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 16/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 17/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 18/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 19/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 20/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 21/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 22/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 23/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 24/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 25/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 26/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 27/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 28/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 29/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 30/1 : 0[3] -> 1[5] via P2P/IPC +ip-26-0-170-160:65607:66197 [3] NCCL INFO Channel 31/1 : 0[3] -> 1[5] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:65608:66202 [4] NCCL INFO Using network Libfabric +ip-26-0-170-160:65610:66203 [6] NCCL INFO Using network Libfabric +ip-26-0-170-160:65610:66203 [6] NCCL INFO comm 0xc096120 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd081859bbfe3dfe9 - Init START +ip-26-0-170-160:65608:66202 [4] NCCL INFO comm 0x13e8cc40 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd081859bbfe3dfe9 - Init START +ip-26-0-170-160:65610:66203 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65608:66202 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65610:66203 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-160:65610:66203 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-160:65610:66203 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-160:65608:66202 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 04/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 05/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 06/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 07/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 08/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 09/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 10/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 11/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 12/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 13/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 14/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 15/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 16/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 17/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 18/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 19/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 20/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 21/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 22/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Channel 23/0 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65608:66202 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65610:66203 [6] NCCL INFO Connected all rings +ip-26-0-170-160:65610:66203 [6] NCCL INFO Connected all trees +ip-26-0-170-160:65610:66203 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65610:66203 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65610:66203 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65608:66202 [4] NCCL INFO Connected all rings +ip-26-0-170-160:65608:66202 [4] NCCL INFO Connected all trees +ip-26-0-170-160:65608:66202 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65608:66202 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65608:66202 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65610:66203 [6] NCCL INFO comm 0xc096120 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd081859bbfe3dfe9 - Init COMPLETE +ip-26-0-170-160:65608:66202 [4] NCCL INFO comm 0x13e8cc40 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xd081859bbfe3dfe9 - Init COMPLETE +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 00/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 01/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 04/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 05/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 06/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 07/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 08/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 09/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 10/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 11/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 12/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 13/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 14/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 15/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 16/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 17/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 18/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 19/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 20/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 21/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 22/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 23/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 24/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 25/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 26/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 27/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 28/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 29/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 30/1 : 0[4] -> 1[6] via P2P/IPC +ip-26-0-170-160:65608:66208 [4] NCCL INFO Channel 31/1 : 0[4] -> 1[6] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:65609:66212 [5] NCCL INFO Using network Libfabric +ip-26-0-170-160:65611:66213 [7] NCCL INFO Using network Libfabric +ip-26-0-170-160:65609:66212 [5] NCCL INFO comm 0x137f4aa0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7ea9ae24ab20945c - Init START +ip-26-0-170-160:65611:66213 [7] NCCL INFO comm 0xb675fa0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7ea9ae24ab20945c - Init START +ip-26-0-170-160:65611:66213 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65609:66212 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-170-160:65611:66213 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-170-160:65611:66213 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-170-160:65611:66213 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-170-160:65609:66212 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 04/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 05/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 06/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 07/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 08/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 09/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 10/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 11/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 12/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 13/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 14/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 15/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 16/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 17/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 18/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 19/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 20/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 21/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 22/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65609:66212 [5] NCCL INFO Channel 23/0 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66213 [7] NCCL INFO Connected all rings +ip-26-0-170-160:65611:66213 [7] NCCL INFO Connected all trees +ip-26-0-170-160:65611:66213 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65611:66213 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65611:66213 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65609:66212 [5] NCCL INFO Connected all rings +ip-26-0-170-160:65609:66212 [5] NCCL INFO Connected all trees +ip-26-0-170-160:65609:66212 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-170-160:65609:66212 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-170-160:65609:66212 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-170-160:65611:66213 [7] NCCL INFO comm 0xb675fa0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7ea9ae24ab20945c - Init COMPLETE +ip-26-0-170-160:65609:66212 [5] NCCL INFO comm 0x137f4aa0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7ea9ae24ab20945c - Init COMPLETE +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 00/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 01/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 04/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 05/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 06/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 07/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 08/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 09/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 10/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 11/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 12/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 13/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 14/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 15/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 16/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 17/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 18/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 19/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 20/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 21/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 22/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 23/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 24/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 25/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 26/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 27/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 28/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 29/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 30/1 : 0[5] -> 1[7] via P2P/IPC +ip-26-0-170-160:65609:66229 [5] NCCL INFO Channel 31/1 : 0[5] -> 1[7] via P2P/IPC +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 160, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 140, in forward + return F.linear(gathered_tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 512.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 331.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 67.57 GiB is allocated by PyTorch, and 229.72 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 284, in train_batch_iter + output = self.forward(context=context, state=state, micro_batch=micro_batch, model=model) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 52, in forward + output = model(**micro_batch) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 1014, in forward + sharded_logits = self.model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 887, in forward + return self.forward_with_hidden_states(input_ids=input_ids, input_mask=input_mask)[0] + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 903, in forward_with_hidden_states + hidden_encoder_states = encoder_block(**hidden_encoder_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 160, in forward + output = self.pp_block(**new_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 752, in forward + hidden_states, sequence_mask = self._core_forward(hidden_states, sequence_mask) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 731, in _core_forward + hidden_states = self.mlp(hidden_states=hidden_states)["hidden_states"] + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 244, in forward + merged_states = self.gate_up_proj(hidden_states) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl + return forward_call(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 89, in forward + return column_linear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 441, in column_linear + return _ColumnLinearAsyncCommunication.apply(input, weight, bias, group, tp_mode, tp_recompute_allgather) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 539, in apply + return super().apply(*args, **kwargs) # type: ignore[misc] + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 140, in forward + return F.linear(gathered_tensor, weight, bias) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 512.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 67.57 GiB is allocated by PyTorch, and 229.72 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +ip-26-0-170-143:2167369:2167821 [6] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167367:2167820 [4] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167365:2167823 [2] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167370:2167714 [7] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167369:2167713 [6] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167368:2167711 [5] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167367:2167712 [4] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167365:2167709 [2] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167366:2167715 [3] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167364:2167708 [1] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167370:2167543 [7] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167368:2167547 [5] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167369:2167544 [6] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167367:2167546 [4] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167365:2167548 [2] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167364:2167545 [1] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167366:2167549 [3] NCCL INFO [Service thread] Connection closed by localRank 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 00/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 01/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 04/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 05/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 06/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 07/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 08/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 09/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 10/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 11/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 12/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 13/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 14/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 15/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 16/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 17/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 18/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 19/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 20/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 21/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 22/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 23/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 24/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 25/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 26/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 27/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 28/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 29/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 30/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-160:65610:66297 [6] NCCL INFO Channel 31/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167368:2167829 [5] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167366:2167830 [3] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167370:2167831 [7] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-170-143:2167370:2167714 [7] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-170-143:2167369:2167713 [6] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-170-143:2167368:2167711 [5] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-170-143:2167367:2167712 [4] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-170-143:2167366:2167715 [3] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-170-143:2167365:2167709 [2] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-170-143:2167370:2167543 [7] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-170-143:2167369:2167544 [6] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-170-143:2167368:2167547 [5] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-170-143:2167367:2167546 [4] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-170-143:2167366:2167549 [3] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-170-143:2167365:2167548 [2] NCCL INFO [Service thread] Connection closed by localRank 1 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 00/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 01/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 04/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 05/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 06/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 07/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 08/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 09/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 10/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 11/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 12/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 13/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 14/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 15/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 16/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 17/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 18/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 19/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 20/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 21/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 22/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 23/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 24/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 25/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 26/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 27/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 28/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 29/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 30/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-160:65611:66300 [7] NCCL INFO Channel 31/1 : 1[7] -> 0[5] via P2P/IPC +[E TraceUtils.h:35] Store is down while updating #7 with key NCCL_0_trace_start +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 00/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 01/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 04/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 05/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 06/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 07/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 08/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 09/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 10/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 11/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 12/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 13/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 14/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 15/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 16/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 17/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 18/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 19/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 20/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 21/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 22/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 23/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 24/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 25/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 26/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 27/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 28/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 29/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 30/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65608:66301 [4] NCCL INFO Channel 31/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 00/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 01/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 04/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 05/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 06/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 07/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 08/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 09/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 10/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 11/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 12/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 13/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 14/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 15/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 16/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 17/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 18/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 19/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 20/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 21/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 22/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 23/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 24/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 25/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 26/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 27/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 28/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 29/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 30/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-160:65609:66303 [5] NCCL INFO Channel 31/1 : 1[5] -> 0[3] via P2P/IPC +[E TraceUtils.h:35] Store is down while updating #10 with key NCCL_0_trace_start +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 00/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 01/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 04/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 05/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 06/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 07/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 08/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 09/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 10/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 11/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 12/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 13/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 14/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 15/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 16/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 17/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 18/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 19/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 20/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 21/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 22/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 23/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 24/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 25/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 26/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 27/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 28/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 29/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 30/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65606:66304 [2] NCCL INFO Channel 31/1 : 1[2] -> 0[0] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 00/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 01/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 04/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 05/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 06/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 07/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 08/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 09/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 10/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 11/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 12/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 13/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 14/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 15/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 16/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 17/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 18/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 19/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 20/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 21/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 22/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 23/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 24/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 25/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 26/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 27/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 28/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 29/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 30/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-160:65607:66306 [3] NCCL INFO Channel 31/1 : 1[3] -> 0[1] via P2P/IPC +ip-26-0-170-143:2167369:2168046 [6] NCCL INFO Channel 02/1 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-143:2167369:2168046 [6] NCCL INFO Channel 03/1 : 1[0] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-170-160:65604:66309 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-160:65604:66309 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[6] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-170-143:2167370:2168048 [7] NCCL INFO Channel 02/1 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-143:2167370:2168048 [7] NCCL INFO Channel 03/1 : 1[1] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-170-160:65605:66310 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-160:65605:66310 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[7] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 00/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 01/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 04/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 05/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 06/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 07/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 08/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 09/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 10/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 11/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 12/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 13/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 14/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 15/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 16/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 17/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 18/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 19/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 20/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 21/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 22/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 23/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 24/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 25/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 26/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 27/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 28/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 29/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 30/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167369:2168051 [6] NCCL INFO Channel 31/1 : 1[6] -> 0[4] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 00/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 01/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 04/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 05/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 06/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 07/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 08/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 09/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 10/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 11/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 12/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 13/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 14/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 15/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 16/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 17/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 18/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 19/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 20/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 21/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 22/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 23/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 24/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 25/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 26/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 27/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 28/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 29/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 30/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167370:2168052 [7] NCCL INFO Channel 31/1 : 1[7] -> 0[5] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 00/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 01/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 04/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 05/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 06/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 07/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 08/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 09/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 10/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 11/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 12/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 13/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 14/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 15/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 16/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 17/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 18/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 19/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 20/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 21/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 22/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 23/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 24/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 25/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 26/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 27/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 28/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 29/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 30/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167367:2168054 [4] NCCL INFO Channel 31/1 : 1[4] -> 0[2] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 00/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 01/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 04/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 05/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 06/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 07/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 08/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 09/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 10/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 11/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 12/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 13/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 14/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 15/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 16/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 17/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 18/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 19/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 20/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 21/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 22/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 23/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 24/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 25/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 26/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 27/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 28/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 29/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 30/1 : 1[5] -> 0[3] via P2P/IPC +ip-26-0-170-143:2167368:2168056 [5] NCCL INFO Channel 31/1 : 1[5] -> 0[3] via P2P/IPC +[2024-12-28 00:54:13,449] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2167364 closing signal SIGTERM +[2024-12-28 00:54:13,449] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2167365 closing signal SIGTERM +[2024-12-28 00:54:13,449] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2167366 closing signal SIGTERM +[2024-12-28 00:54:13,449] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2167367 closing signal SIGTERM +[2024-12-28 00:54:13,449] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2167368 closing signal SIGTERM +[2024-12-28 00:54:13,449] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2167369 closing signal SIGTERM +[2024-12-28 00:54:13,449] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2167370 closing signal SIGTERM +[2024-12-28 00:54:43,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Unable to shutdown process 2167364 via Signals.SIGTERM, forcefully exiting via Signals.SIGKILL +12/28/2024 01:14:07 [WARNING|DP=0|PP=4|TP=0|ip-26-0-170-160]: [RECV META] Rank 4: First metadata receive from rank 3 timed out after 1200s +12/28/2024 01:14:07 [ERROR|DP=0|PP=4|TP=0|ip-26-0-170-160]: [RECV META] Rank 4: Failed to receive metadata from rank 3: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:07 [WARNING|DP=0|PP=4|TP=0|ip-26-0-170-160]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:08 [WARNING|DP=0|PP=5|TP=0|ip-26-0-170-160]: [RECV META] Rank 5: First metadata receive from rank 4 timed out after 1200s +12/28/2024 01:14:08 [ERROR|DP=0|PP=5|TP=0|ip-26-0-170-160]: [RECV META] Rank 5: Failed to receive metadata from rank 4: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:08 [WARNING|DP=0|PP=5|TP=0|ip-26-0-170-160]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:08 [WARNING|DP=1|PP=4|TP=0|ip-26-0-170-160]: [RECV META] Rank 4: First metadata receive from rank 3 timed out after 1200s +12/28/2024 01:14:08 [ERROR|DP=1|PP=4|TP=0|ip-26-0-170-160]: [RECV META] Rank 4: Failed to receive metadata from rank 3: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:08 [WARNING|DP=1|PP=4|TP=0|ip-26-0-170-160]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 3: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:08 [WARNING|DP=0|PP=7|TP=0|ip-26-0-170-160]: [RECV META] Rank 7: First metadata receive from rank 6 timed out after 1200s +12/28/2024 01:14:08 [ERROR|DP=0|PP=7|TP=0|ip-26-0-170-160]: [RECV META] Rank 7: Failed to receive metadata from rank 6: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:08 [WARNING|DP=0|PP=7|TP=0|ip-26-0-170-160]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:09 [WARNING|DP=0|PP=6|TP=0|ip-26-0-170-160]: [RECV META] Rank 6: First metadata receive from rank 5 timed out after 1200s +12/28/2024 01:14:09 [ERROR|DP=0|PP=6|TP=0|ip-26-0-170-160]: [RECV META] Rank 6: Failed to receive metadata from rank 5: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:09 [WARNING|DP=0|PP=6|TP=0|ip-26-0-170-160]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:09 [WARNING|DP=1|PP=5|TP=0|ip-26-0-170-160]: [RECV META] Rank 5: First metadata receive from rank 4 timed out after 1200s +12/28/2024 01:14:09 [ERROR|DP=1|PP=5|TP=0|ip-26-0-170-160]: [RECV META] Rank 5: Failed to receive metadata from rank 4: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:09 [WARNING|DP=1|PP=5|TP=0|ip-26-0-170-160]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 4: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:09 [WARNING|DP=1|PP=7|TP=0|ip-26-0-170-160]: [RECV META] Rank 7: First metadata receive from rank 6 timed out after 1200s +12/28/2024 01:14:09 [ERROR|DP=1|PP=7|TP=0|ip-26-0-170-160]: [RECV META] Rank 7: Failed to receive metadata from rank 6: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:09 [WARNING|DP=1|PP=7|TP=0|ip-26-0-170-160]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 6: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:09 [WARNING|DP=1|PP=6|TP=0|ip-26-0-170-160]: [RECV META] Rank 6: First metadata receive from rank 5 timed out after 1200s +12/28/2024 01:14:09 [ERROR|DP=1|PP=6|TP=0|ip-26-0-170-160]: [RECV META] Rank 6: Failed to receive metadata from rank 5: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +12/28/2024 01:14:09 [WARNING|DP=1|PP=6|TP=0|ip-26-0-170-160]: recv_activation failed, will retry later. Error: Failed to receive metadata from rank 5: 'torch._C._distributed_c10d.Work' object has no attribute 'abort' +slurmstepd: error: *** JOB 13848169 ON ip-26-0-170-143 CANCELLED AT 2024-12-28T02:02:51 DUE TO TIME LIMIT *** +--- Logging error --- +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +slurmstepd: error: *** STEP 13848169.0 ON ip-26-0-170-143 CANCELLED AT 2024-12-28T02:02:51 DUE TO TIME LIMIT *** +[2024-12-28 02:02:51,282] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-28 02:02:51,287] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2167364 closing signal SIGTERM +[2024-12-28 02:02:51,288] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2167368 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 65533 got signal: 15 + +During handling of the above exception, another exception occurred: + +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/logging/__init__.py", line 1100, in emit + msg = self.format(record) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/logging/__init__.py", line 943, in format + return fmt.format(record) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_logging/_internal.py", line 635, in format + record.message = record.getMessage() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/logging/__init__.py", line 368, in getMessage + msg = msg % self.args + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 65533 got signal: 15 +Call stack: + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 742, in run + log.warning("Received %s death signal, shutting down workers", e.sigval) +Message: 'Received %s death signal, shutting down workers' +Arguments: (,) +[2024-12-28 02:02:51,293] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 65604 closing signal SIGTERM +[2024-12-28 02:02:51,294] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 65605 closing signal SIGTERM +[2024-12-28 02:02:51,294] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 65606 closing signal SIGTERM +[2024-12-28 02:02:51,294] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 65607 closing signal SIGTERM +[2024-12-28 02:02:51,294] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 65608 closing signal SIGTERM +[2024-12-28 02:02:51,294] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 65609 closing signal SIGTERM +[2024-12-28 02:02:51,294] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 65610 closing signal SIGTERM +[2024-12-28 02:02:51,294] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 65611 closing signal SIGTERM +[2024-12-28 02:02:51,309] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2167364 closing signal SIGTERM +[2024-12-28 02:02:51,310] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2167368 closing signal SIGTERM diff --git a/logs/13848317-bench_1.34G_dp8_tp1_pp2_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13848317-bench_1.34G_dp8_tp1_pp2_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..d1fc1c07c39661471a60940f09db268ba02f0c3c --- /dev/null +++ b/logs/13848317-bench_1.34G_dp8_tp1_pp2_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,6887 @@ ++ '[' -z 13848317 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames ip-26-0-168-52,ip-26-0-171-21 ++ export 'NODELIST=ip-26-0-168-52 +ip-26-0-171-21' ++ NODELIST='ip-26-0-168-52 +ip-26-0-171-21' +++ scontrol show hostnames ip-26-0-168-52,ip-26-0-171-21 +++ head -n1 ++ export MASTER_NODE=ip-26-0-168-52 ++ MASTER_NODE=ip-26-0-168-52 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=2 ++ NNODES=2 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=16 ++ WORLD_SIZE=16 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-168-52' +Master node: ip-26-0-168-52 ++ echo 'All nodes: ip-26-0-168-52 +ip-26-0-171-21' +All nodes: ip-26-0-168-52 +ip-26-0-171-21 ++ echo 'World size: 16' +World size: 16 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=2 --nproc_per_node=8 --rdzv_id=13848317 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-168-52:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.34G_dp8_tp1_pp2_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-28 01:07:11,975] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 01:07:11,975] torch.distributed.run: [WARNING] +[2024-12-28 01:07:11,975] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 01:07:11,975] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 01:07:11,975] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 01:07:12,075] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 01:07:12,075] torch.distributed.run: [WARNING] +[2024-12-28 01:07:12,075] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 01:07:12,075] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 01:07:12,075] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-168-52:2509869:2509869 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-52:2509869:2509869 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.168.52<0> +ip-26-0-168-52:2509869:2509869 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-52:2509869:2509869 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-52:2509869:2509869 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-171-21:1183730:1183730 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-171-21:1183731:1183731 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-52:2509873:2509873 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-52:2509870:2509870 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-52:2509871:2509871 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-171-21:1183730:1183730 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-52:2509875:2509875 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-171-21:1183731:1183731 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-52:2509870:2509870 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-52:2509873:2509873 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-52:2509871:2509871 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-52:2509876:2509876 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-52:2509874:2509874 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-52:2509875:2509875 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-171-21:1183727:1183727 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-52:2509876:2509876 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-52:2509874:2509874 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-171-21:1183727:1183727 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-171-21:1183730:1183730 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.171.21<0> +ip-26-0-171-21:1183731:1183731 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.171.21<0> +ip-26-0-168-52:2509872:2509872 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-52:2509872:2509872 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-171-21:1183727:1183727 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.171.21<0> +ip-26-0-171-21:1183730:1183730 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-171-21:1183730:1183730 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-171-21:1183731:1183731 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-171-21:1183731:1183731 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-52:2509870:2509870 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.168.52<0> +ip-26-0-168-52:2509873:2509873 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.168.52<0> +ip-26-0-168-52:2509871:2509871 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.168.52<0> +ip-26-0-168-52:2509875:2509875 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.168.52<0> +ip-26-0-171-21:1183727:1183727 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-171-21:1183727:1183727 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-52:2509874:2509874 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.168.52<0> +ip-26-0-168-52:2509876:2509876 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.168.52<0> +ip-26-0-168-52:2509870:2509870 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-52:2509873:2509873 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-52:2509871:2509871 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-52:2509870:2509870 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-52:2509873:2509873 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-52:2509871:2509871 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-52:2509875:2509875 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-52:2509875:2509875 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-52:2509872:2509872 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.168.52<0> +ip-26-0-168-52:2509874:2509874 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-52:2509874:2509874 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-52:2509872:2509872 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-52:2509872:2509872 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-52:2509876:2509876 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-52:2509876:2509876 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-171-21:1183733:1183733 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-171-21:1183728:1183728 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-171-21:1183734:1183734 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-171-21:1183729:1183729 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-171-21:1183732:1183732 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-171-21:1183733:1183733 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-171-21:1183728:1183728 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-171-21:1183734:1183734 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-171-21:1183729:1183729 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-171-21:1183732:1183732 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-171-21:1183733:1183733 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.171.21<0> +ip-26-0-171-21:1183728:1183728 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.171.21<0> +ip-26-0-171-21:1183732:1183732 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.171.21<0> +ip-26-0-171-21:1183729:1183729 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.171.21<0> +ip-26-0-171-21:1183734:1183734 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.171.21<0> +ip-26-0-171-21:1183732:1183732 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-171-21:1183733:1183733 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-171-21:1183732:1183732 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-171-21:1183733:1183733 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-171-21:1183729:1183729 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-171-21:1183729:1183729 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-171-21:1183728:1183728 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-171-21:1183728:1183728 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-171-21:1183734:1183734 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-171-21:1183734:1183734 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/205 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/184 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO comm 0x98b1be0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO comm 0x986a520 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO comm 0x8c18260 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO comm 0x9011ca0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO comm 0x9ed5620 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO comm 0x99df5b0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO comm 0x97c5700 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO comm 0x843d430 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO comm 0xa3b65b0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO comm 0x98ca420 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO comm 0x8851530 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO comm 0x8838840 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO comm 0x978ecf0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO comm 0x92fe750 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO comm 0x85af2e0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO comm 0x98ae7d0 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1bfd7e2d255dcd8d - Init START +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->1 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/1/-1->9->-1 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/14/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->2 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/2/-1->10->-1 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->7 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/7/-1->15->-1 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->6 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/6/-1->14->-1 [15] -1/-1/-1->14->13 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/15/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Trees [0] 9/-1/-1->8->0 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/0/-1->8->-1 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->5 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/5/-1->13->-1 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/11/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->3 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/3/-1->11->-1 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->4 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/4/-1->12->-1 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/12/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 01/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 1 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 2 1 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 03/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 3 2 1 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 04/16 : 0 7 6 5 12 11 10 9 8 15 14 13 4 3 2 1 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 05/16 : 0 4 7 6 13 11 10 9 8 12 15 14 5 3 2 1 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/9/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 06/16 : 0 5 4 7 14 11 10 9 8 13 12 15 6 3 2 1 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 07/16 : 0 6 5 4 15 11 10 9 8 14 13 12 7 3 2 1 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/13/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 09/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 1 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 2 1 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 11/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 3 2 1 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 12/16 : 0 7 6 5 12 11 10 9 8 15 14 13 4 3 2 1 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 13/16 : 0 4 7 6 13 11 10 9 8 12 15 14 5 3 2 1 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 14/16 : 0 5 4 7 14 11 10 9 8 13 12 15 6 3 2 1 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 15/16 : 0 6 5 4 15 11 10 9 8 14 13 12 7 3 2 1 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Trees [0] 1/8/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/10/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 05/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 06/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 05/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 13/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 06/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 13/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 14/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 14/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 08/0 : 15[7] -> 0[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 08/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 02/0 : 11[3] -> 2[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 10/0 : 11[3] -> 2[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 05/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 13/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 06/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 06/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 14/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 14/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 07/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 07/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 15/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 02/0 : 11[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 10/0 : 11[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 15/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 06/0 : 15[7] -> 6[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 03/0 : 12[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 11/0 : 12[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 01/0 : 10[2] -> 1[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 09/0 : 10[2] -> 1[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 03/0 : 12[4] -> 3[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 01/0 : 10[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 11/0 : 12[4] -> 3[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 09/0 : 10[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 14/0 : 15[7] -> 6[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 05/0 : 14[6] -> 5[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 13/0 : 14[6] -> 5[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 04/0 : 13[5] -> 4[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 12/0 : 13[5] -> 4[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 04/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 05/0 : 14[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 12/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 13/0 : 14[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 06/0 : 15[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 14/0 : 15[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 06/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 05/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 06/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 05/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 14/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 13/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 14/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 13/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 07/0 : 12[4] -> 7[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 10/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 09/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 15/0 : 12[4] -> 7[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 07/0 : 12[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 15/0 : 12[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 07/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 15/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 12/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 11/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Connected all rings +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Connected all rings +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Connected all rings +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Connected all rings +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Connected all rings +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Connected all rings +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Connected all rings +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Connected all rings +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Connected all rings +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Connected all rings +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Connected all rings +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Connected all rings +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Connected all rings +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Connected all trees +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NVLS comm 0x92fe750 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Connected all trees +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NVLS comm 0x85af2e0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Connected all trees +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NVLS comm 0x98ca420 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Connected all trees +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Connected all trees +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NVLS comm 0x978ecf0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NVLS comm 0x98ae7d0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Connected all trees +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Connected all trees +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Connected all trees +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NVLS comm 0x8838840 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Connected all trees +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NVLS comm 0x97c5700 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NVLS comm 0x9ed5620 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NVLS comm 0x8c18260 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Connected all trees +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Connected all trees +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NVLS comm 0x8851530 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NVLS comm 0x9011ca0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Connected all trees +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NVLS comm 0x99df5b0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Connected all trees +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NVLS comm 0x98b1be0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Connected all trees +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NVLS comm 0xa3b65b0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Connected all trees +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NVLS comm 0x986a520 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Connected all trees +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NVLS comm 0x843d430 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183728:1183893 [1] NCCL INFO comm 0xa3b65b0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-171-21:1183734:1183894 [7] NCCL INFO comm 0x8838840 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-171-21:1183732:1183891 [5] NCCL INFO comm 0x92fe750 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-171-21:1183727:1183887 [0] NCCL INFO comm 0x978ecf0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-171-21:1183733:1183890 [6] NCCL INFO comm 0x8851530 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-171-21:1183730:1183889 [3] NCCL INFO comm 0x98ae7d0 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-171-21:1183731:1183888 [4] NCCL INFO comm 0x85af2e0 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-171-21:1183729:1183892 [2] NCCL INFO comm 0x98ca420 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-168-52:2509874:2510032 [5] NCCL INFO comm 0x9011ca0 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-168-52:2509870:2510036 [1] NCCL INFO comm 0x843d430 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-168-52:2509875:2510037 [6] NCCL INFO comm 0x986a520 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-168-52:2509873:2510033 [4] NCCL INFO comm 0x97c5700 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-168-52:2509869:2510030 [0] NCCL INFO comm 0x99df5b0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-168-52:2509871:2510034 [2] NCCL INFO comm 0x9ed5620 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-168-52:2509872:2510035 [3] NCCL INFO comm 0x8c18260 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-168-52:2509876:2510031 [7] NCCL INFO comm 0x98b1be0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1bfd7e2d255dcd8d - Init COMPLETE +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO comm 0x952c4c0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO comm 0x8a67040 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO comm 0x8a7f440 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO comm 0x9adc980 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO comm 0x87dd820 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO comm 0x9af81e0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO comm 0x9ae00c0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO comm 0x9a98010 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO comm 0x9240280 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO comm 0x99f4950 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO comm 0x9c116c0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO comm 0x8e46d80 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO comm 0xa5e48a0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO comm 0x99bbde0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO comm 0x866b580 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO comm 0xa103790 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6cfd7c371d241d8a - Init START +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/10/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->6 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/6/-1->14->-1 [15] -1/-1/-1->14->13 +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->5 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/5/-1->13->-1 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->7 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/7/-1->15->-1 +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->2 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/2/-1->10->-1 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 01/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 1 +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->3 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/3/-1->11->-1 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->4 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/4/-1->12->-1 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 2 1 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 03/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 3 2 1 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/9/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 04/16 : 0 7 6 5 12 11 10 9 8 15 14 13 4 3 2 1 +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->1 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/1/-1->9->-1 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Trees [0] 9/-1/-1->8->0 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/0/-1->8->-1 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 05/16 : 0 4 7 6 13 11 10 9 8 12 15 14 5 3 2 1 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 06/16 : 0 5 4 7 14 11 10 9 8 13 12 15 6 3 2 1 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 07/16 : 0 6 5 4 15 11 10 9 8 14 13 12 7 3 2 1 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 09/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 1 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 2 1 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 11/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 3 2 1 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 12/16 : 0 7 6 5 12 11 10 9 8 15 14 13 4 3 2 1 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NVLS Head 0: 0 8 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 13/16 : 0 4 7 6 13 11 10 9 8 12 15 14 5 3 2 1 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 14/16 : 0 5 4 7 14 11 10 9 8 13 12 15 6 3 2 1 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NVLS Head 1: 1 9 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NVLS Head 2: 2 10 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 15/16 : 0 6 5 4 15 11 10 9 8 14 13 12 7 3 2 1 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NVLS Head 3: 3 11 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/11/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NVLS Head 4: 4 12 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NVLS Head 5: 5 13 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Trees [0] 1/8/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NVLS Head 6: 6 14 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NVLS Head 7: 7 15 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/15/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/14/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/12/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/13/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 05/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 05/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 06/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 13/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 14/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 06/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 13/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 14/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 00/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 08/0 : 15[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 08/0 : 15[7] -> 0[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 02/0 : 11[3] -> 2[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 10/0 : 11[3] -> 2[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 05/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 01/0 : 10[2] -> 1[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 13/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 06/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 14/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 03/0 : 12[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 07/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 11/0 : 12[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 06/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 15/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 02/0 : 11[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 14/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 10/0 : 11[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 07/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 15/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 06/0 : 15[7] -> 6[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 14/0 : 15[7] -> 6[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 09/0 : 10[2] -> 1[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 01/0 : 10[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 09/0 : 10[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 03/0 : 12[4] -> 3[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 11/0 : 12[4] -> 3[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 04/0 : 13[5] -> 4[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 12/0 : 13[5] -> 4[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 04/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 12/0 : 13[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 10/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 06/0 : 15[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 14/0 : 15[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 09/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 05/0 : 14[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 13/0 : 14[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 06/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 05/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 05/0 : 14[6] -> 5[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 14/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 13/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 13/0 : 14[6] -> 5[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 05/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 06/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 14/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 13/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 07/0 : 12[4] -> 7[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 15/0 : 12[4] -> 7[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 07/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 07/0 : 12[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 15/0 : 12[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 15/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 12/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 11/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Connected all rings +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Connected all rings +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Connected all rings +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Connected all rings +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Connected all rings +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Connected all rings +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Connected all rings +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Connected all rings +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Connected all rings +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Connected all rings +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 04/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 04/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Connected all rings +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 00/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 00/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 02/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 02/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 06/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 06/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Connected all trees +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO NVLS comm 0x99bbde0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Connected all trees +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO NVLS comm 0x8a67040 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Connected all trees +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NVLS comm 0x9c116c0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Connected all trees +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO NVLS comm 0x952c4c0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Connected all trees +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO NVLS comm 0x8a7f440 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Connected all trees +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO NVLS comm 0xa5e48a0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Connected all trees +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO NVLS comm 0x9adc980 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Connected all trees +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Connected all trees +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NVLS comm 0x9ae00c0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NVLS comm 0x9240280 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Connected all trees +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Connected all trees +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Connected all trees +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NVLS comm 0x9a98010 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO NVLS comm 0x87dd820 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO NVLS comm 0x9af81e0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Connected all trees +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NVLS comm 0x866b580 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Connected all trees +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Connected all trees +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NVLS comm 0x99f4950 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NVLS comm 0x8e46d80 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Connected all trees +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NVLS comm 0xa103790 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 00/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 00/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 00/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 00/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 02/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 02/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 02/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 02/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 04/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 04/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 04/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 06/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 04/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 06/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 06/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 08/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 06/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 08/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 10/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 08/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 08/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 10/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 12/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 10/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 12/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 10/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 14/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 12/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 14/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 12/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 00/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 14/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 00/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 14/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 02/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 02/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 00/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 00/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 04/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 04/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 02/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 06/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 02/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 04/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 06/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 04/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 08/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 06/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 08/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 10/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 10/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 06/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 08/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 08/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 12/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 12/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 14/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 10/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 10/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 12/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 14/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 12/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 14/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Channel 14/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 02/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 04/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 06/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 10/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 12/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 14/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 02/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 04/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 06/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 10/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 12/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 14/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 00/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 02/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 06/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 08/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 10/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 00/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 14/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 00/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 00/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 04/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 02/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 02/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 04/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 06/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 06/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 08/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 08/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 08/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 10/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 10/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 12/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 12/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 14/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 14/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 00/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 00/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 02/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 04/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 04/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 06/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 08/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 08/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 10/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 12/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 12/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 14/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO threadThresholds 8/8/64 | 128/8/64 | 512 | 512 +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509874:2510172 [5] NCCL INFO comm 0x9240280 rank 5 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-168-52:2509876:2510175 [7] NCCL INFO comm 0x9ae00c0 rank 7 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-168-52:2509872:2510170 [3] NCCL INFO comm 0x8e46d80 rank 3 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-168-52:2509870:2510169 [1] NCCL INFO comm 0x866b580 rank 1 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-168-52:2509873:2510173 [4] NCCL INFO comm 0x99f4950 rank 4 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-168-52:2509869:2510168 [0] NCCL INFO comm 0x9c116c0 rank 0 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-168-52:2509871:2510171 [2] NCCL INFO comm 0xa103790 rank 2 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-168-52:2509875:2510174 [6] NCCL INFO comm 0x9a98010 rank 6 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-171-21:1183730:1184027 [3] NCCL INFO comm 0x9adc980 rank 11 nranks 16 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-171-21:1183732:1184022 [5] NCCL INFO comm 0x952c4c0 rank 13 nranks 16 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-171-21:1183728:1184025 [1] NCCL INFO comm 0xa5e48a0 rank 9 nranks 16 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-171-21:1183734:1184028 [7] NCCL INFO comm 0x8a67040 rank 15 nranks 16 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-171-21:1183729:1184026 [2] NCCL INFO comm 0x9af81e0 rank 10 nranks 16 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-171-21:1183727:1184029 [0] NCCL INFO comm 0x99bbde0 rank 8 nranks 16 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-171-21:1183731:1184024 [4] NCCL INFO comm 0x87dd820 rank 12 nranks 16 cudaDev 4 nvmlDev 4 busId 97000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-171-21:1183733:1184023 [6] NCCL INFO comm 0x8a7f440 rank 14 nranks 16 cudaDev 6 nvmlDev 6 busId b9000 commId 0x6cfd7c371d241d8a - Init COMPLETE +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO comm 0x8e5db90 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa3d929a70e0e66d5 - Init START +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO comm 0x9a0b9a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa3d929a70e0e66d5 - Init START +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO comm 0x9256a40 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa3d929a70e0e66d5 - Init START +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO comm 0x9c27c70 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa3d929a70e0e66d5 - Init START +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO comm 0x9aad8c0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa3d929a70e0e66d5 - Init START +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO comm 0xa11a750 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa3d929a70e0e66d5 - Init START +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO comm 0x9af6b90 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa3d929a70e0e66d5 - Init START +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO comm 0x8682f90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa3d929a70e0e66d5 - Init START +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO comm 0x9b0e1c0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xbbdedd9938e3b5f5 - Init START +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO comm 0xa5fbac0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbbdedd9938e3b5f5 - Init START +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO comm 0x9af43d0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbbdedd9938e3b5f5 - Init START +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO comm 0x8a95160 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbbdedd9938e3b5f5 - Init START +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO comm 0x9542be0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbbdedd9938e3b5f5 - Init START +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO comm 0x8a7e3d0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbbdedd9938e3b5f5 - Init START +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO comm 0x99d4ad0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xbbdedd9938e3b5f5 - Init START +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO comm 0x87f3220 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xbbdedd9938e3b5f5 - Init START +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Connected all rings +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Connected all rings +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Connected all rings +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Connected all rings +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Connected all rings +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Connected all rings +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Connected all rings +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Connected all rings +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Connected all trees +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO NVLS comm 0x99d4ad0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Connected all rings +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Connected all rings +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Connected all rings +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Connected all rings +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Connected all rings +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Connected all rings +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Connected all rings +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Connected all trees +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO NVLS comm 0x8a7e3d0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Connected all trees +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO NVLS comm 0xa5fbac0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Connected all trees +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO NVLS comm 0x9c27c70 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Connected all trees +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO NVLS comm 0x8a95160 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Connected all trees +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO NVLS comm 0x9af43d0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Connected all trees +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO NVLS comm 0x87f3220 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Connected all trees +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO NVLS comm 0x9542be0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Connected all trees +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO NVLS comm 0x9b0e1c0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO Connected NVLS tree +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183733:1184060 [6] NCCL INFO comm 0x8a95160 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xbbdedd9938e3b5f5 - Init COMPLETE +ip-26-0-171-21:1183727:1184058 [0] NCCL INFO comm 0x99d4ad0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xbbdedd9938e3b5f5 - Init COMPLETE +ip-26-0-171-21:1183729:1184062 [2] NCCL INFO comm 0x9b0e1c0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xbbdedd9938e3b5f5 - Init COMPLETE +ip-26-0-171-21:1183732:1184059 [5] NCCL INFO comm 0x9542be0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xbbdedd9938e3b5f5 - Init COMPLETE +ip-26-0-171-21:1183731:1184063 [4] NCCL INFO comm 0x87f3220 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xbbdedd9938e3b5f5 - Init COMPLETE +ip-26-0-171-21:1183730:1184061 [3] NCCL INFO comm 0x9af43d0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbbdedd9938e3b5f5 - Init COMPLETE +ip-26-0-171-21:1183734:1184064 [7] NCCL INFO comm 0x8a7e3d0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xbbdedd9938e3b5f5 - Init COMPLETE +ip-26-0-171-21:1183728:1184065 [1] NCCL INFO comm 0xa5fbac0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xbbdedd9938e3b5f5 - Init COMPLETE +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Connected all trees +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO NVLS comm 0x9af6b90 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Connected all trees +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO NVLS comm 0x9a0b9a0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Connected all trees +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO NVLS comm 0xa11a750 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Connected all trees +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO NVLS comm 0x9256a40 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Connected all trees +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO NVLS comm 0x8682f90 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Connected all trees +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO NVLS comm 0x8e5db90 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Connected all trees +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO NVLS comm 0x9aad8c0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO Connected NVLS tree +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509876:2510205 [7] NCCL INFO comm 0x9af6b90 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa3d929a70e0e66d5 - Init COMPLETE +ip-26-0-168-52:2509869:2510202 [0] NCCL INFO comm 0x9c27c70 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa3d929a70e0e66d5 - Init COMPLETE +ip-26-0-168-52:2509871:2510209 [2] NCCL INFO comm 0xa11a750 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa3d929a70e0e66d5 - Init COMPLETE +ip-26-0-168-52:2509875:2510204 [6] NCCL INFO comm 0x9aad8c0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa3d929a70e0e66d5 - Init COMPLETE +ip-26-0-168-52:2509873:2510207 [4] NCCL INFO comm 0x9a0b9a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa3d929a70e0e66d5 - Init COMPLETE +ip-26-0-168-52:2509872:2510208 [3] NCCL INFO comm 0x8e5db90 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa3d929a70e0e66d5 - Init COMPLETE +ip-26-0-168-52:2509874:2510203 [5] NCCL INFO comm 0x9256a40 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa3d929a70e0e66d5 - Init COMPLETE +ip-26-0-168-52:2509870:2510206 [1] NCCL INFO comm 0x8682f90 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa3d929a70e0e66d5 - Init COMPLETE +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Config: +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Config(general=GeneralArgs(project='debug', +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: run='1.34G_dp8_tp1_pp2_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k', +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: seed=42, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: step=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: consumed_train_samples=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: ignore_sanity_checks=True), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: parallelism=ParallelismArgs(dp=8, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: pp=2, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: tp=1, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: pp_engine=, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: tp_mode=, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: tp_linear_async_communication=True, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: recompute_layer=False, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: tp_recompute_allgather=True, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: expert_parallel_size=1), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: eos_token_id=0, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: hidden_act='silu', +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: hidden_size=2048, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: initializer_range=0.02, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: intermediate_size=8192, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: is_llama_config=True, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: max_position_embeddings=4096, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: num_attention_heads=32, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: num_hidden_layers=16, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: num_key_value_heads=32, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: pad_token_id=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: pretraining_tp=1, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: rms_norm_eps=1e-05, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: rope_scaling=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: rope_theta=10000.0, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: rope_interleaved=False, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: tie_word_embeddings=True, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: use_cache=True, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: vocab_size=131072), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: init_method=RandomInit(std=0.02), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: dtype=torch.bfloat16, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: make_vocab_size_divisible_by=1, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: ddp_bucket_cap_mb=25), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: tokenizer_revision=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: tokenizer_max_length=None), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: checkpoint_interval=10000, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: save_initial_state=False, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: save_final_state=False, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: resume_checkpoint_path=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: checkpoints_path_is_shared_file_system=False), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: logging=LoggingArgs(log_level='info', +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: log_level_replica='info', +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: iteration_step_info_interval=1), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: tokens=TokensArgs(sequence_length=4096, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: train_steps=100, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: micro_batch_size=1, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: batch_accumulation_per_replica=32, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: val_check_interval=100, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: limit_val_batches=0, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: limit_test_batches=0), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: adam_beta1=0.9, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: adam_beta2=0.95, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: torch_adam_is_fused=True, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: name='adamW'), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: zero_stage=1, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: weight_decay=0.01, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: clip_grad=1.0, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: accumulate_grad_in_fp32=True, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: lr_warmup_steps=2, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: lr_warmup_style='linear', +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: lr_decay_style='cosine', +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: lr_decay_steps=13, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: lr_decay_starting_step=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: min_decay_lr=1e-05)), +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: start_training_step=1, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: data=DataArgs(dataset=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: seed=42, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: num_loading_workers=1))], +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: profiler=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: lighteval=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: s3_upload=None) +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Model Config: +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: LlamaConfig(bos_token_id=0, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: eos_token_id=0, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: hidden_act='silu', +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: hidden_size=2048, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: initializer_range=0.02, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: intermediate_size=8192, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: is_llama_config=True, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: max_position_embeddings=4096, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: num_attention_heads=32, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: num_hidden_layers=16, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: num_key_value_heads=32, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: pad_token_id=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: pretraining_tp=1, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: rms_norm_eps=1e-05, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: rope_scaling=None, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: rope_theta=10000.0, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: rope_interleaved=False, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: tie_word_embeddings=True, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: use_cache=True, +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: vocab_size=131072) +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Building model.. +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Initialize RoPE Theta = 10000.0 +12/28/2024 01:07:50 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO comm 0xa6fcfb0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd260c33e8fe4b5b - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO comm 0xb775ef0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x685c1dacbad34513 - Init START +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO comm 0xb1ab6b0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1e5af083178c9f86 - Init START +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO comm 0xa45acd0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf14110c4145c5852 - Init START +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO comm 0xb63eb30 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe0352c6a6399204c - Init START +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO comm 0xb08a600 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe5e172062bb4e3bb - Init START +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO comm 0xc2630b0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x936c927d48399cda - Init START +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO comm 0xac915c0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x577054749f6c445d - Init START +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO comm 0xb75c000 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe89f1bf8ed0edb3c - Init START +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO comm 0xb92b1b0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3dfecec0ea36353e - Init START +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO comm 0xb83f130 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc8ec4a2bf8e646af - Init START +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO comm 0xb8e52b0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x38c629418b2965c1 - Init START +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO comm 0xa6e4c30 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98264197a51f37a - Init START +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO comm 0xbf4cbb0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe0f39be28062b4f8 - Init START +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO comm 0xba610d0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5ffe66ef3933e525 - Init START +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO comm 0xa4b7df0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe04fd2999db3a8de - Init START +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Connected all rings +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO Connected all trees +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Connected all rings +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO Connected all trees +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183727:1184099 [0] NCCL INFO comm 0xb63eb30 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe0352c6a6399204c - Init COMPLETE +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Connected all rings +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO Connected all trees +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Connected all rings +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO Connected all trees +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183730:1184105 [3] NCCL INFO comm 0xb75c000 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe89f1bf8ed0edb3c - Init COMPLETE +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Connected all rings +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO Connected all trees +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO Connected all trees +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Connected all rings +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO Connected all trees +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183734:1184107 [7] NCCL INFO comm 0xa6e4c30 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98264197a51f37a - Init COMPLETE +ip-26-0-168-52:2509874:2510238 [5] NCCL INFO comm 0xb08a600 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe5e172062bb4e3bb - Init COMPLETE +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Connected all rings +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO Connected all trees +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Connected all rings +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO Connected all trees +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Connected all rings +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO Connected all trees +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183729:1184103 [2] NCCL INFO comm 0xb775ef0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x685c1dacbad34513 - Init COMPLETE +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Connected all rings +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO Connected all trees +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-171-21:1183731:1184100 [4] NCCL INFO comm 0xa45acd0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf14110c4145c5852 - Init COMPLETE +ip-26-0-171-21:1183733:1184096 [6] NCCL INFO comm 0xa6fcfb0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd260c33e8fe4b5b - Init COMPLETE +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Connected all rings +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO Connected all trees +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509870:2510248 [1] NCCL INFO comm 0xa4b7df0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xe04fd2999db3a8de - Init COMPLETE +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Connected all rings +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO Connected all trees +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183728:1184104 [1] NCCL INFO comm 0xc2630b0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x936c927d48399cda - Init COMPLETE +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Connected all rings +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO Connected all trees +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509869:2510251 [0] NCCL INFO comm 0xba610d0 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5ffe66ef3933e525 - Init COMPLETE +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Connected all rings +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO Connected all trees +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183732:1184098 [5] NCCL INFO comm 0xb1ab6b0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1e5af083178c9f86 - Init COMPLETE +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO comm 0xc265d60 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeb22685fc134aa4a - Init START +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO comm 0xa4baaa0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeb22685fc134aa4a - Init START +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO comm 0xb6417e0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf1eadbc30220ebb9 - Init START +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO comm 0xba63d80 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf1eadbc30220ebb9 - Init START +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO comm 0xb1ae360 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd3eba80527af902 - Init START +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO comm 0xb08d2b0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd3eba80527af902 - Init START +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Connected all rings +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO Connected all trees +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-52:2509871:2510250 [2] NCCL INFO comm 0xbf4cbb0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xe0f39be28062b4f8 - Init COMPLETE +ip-26-0-168-52:2509876:2510242 [7] NCCL INFO comm 0xb92b1b0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3dfecec0ea36353e - Init COMPLETE +ip-26-0-168-52:2509873:2510243 [4] NCCL INFO comm 0xb83f130 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc8ec4a2bf8e646af - Init COMPLETE +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO comm 0xb778ba0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd8cd3a80175e00f7 - Init START +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO comm 0xbf4f860 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd8cd3a80175e00f7 - Init START +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509875:2510245 [6] NCCL INFO comm 0xb8e52b0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x38c629418b2965c1 - Init COMPLETE +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO comm 0xa6e78e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56724530d017e8f1 - Init START +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO comm 0xb92de60 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56724530d017e8f1 - Init START +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO comm 0xa45d980 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4bf8d575fb167966 - Init START +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO comm 0xb841de0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4bf8d575fb167966 - Init START +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO comm 0xa6ffc60 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x866e037d1ddeeb05 - Init START +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO comm 0xb8e7f60 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x866e037d1ddeeb05 - Init START +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509872:2510244 [3] NCCL INFO comm 0xac915c0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x577054749f6c445d - Init COMPLETE +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Using network Libfabric +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO comm 0xac94270 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfe7d2120cde959e - Init START +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO comm 0xb75ecb0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfe7d2120cde959e - Init START +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Connected all rings +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO Connected all trees +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Connected all rings +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO Connected all trees +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183732:1184126 [5] NCCL INFO comm 0xb1ae360 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd3eba80527af902 - Init COMPLETE +ip-26-0-168-52:2509874:2510272 [5] NCCL INFO comm 0xb08d2b0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xd3eba80527af902 - Init COMPLETE +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Connected all rings +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO Connected all trees +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Connected all rings +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO Connected all trees +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509870:2510266 [1] NCCL INFO comm 0xa4baaa0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeb22685fc134aa4a - Init COMPLETE +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Connected all rings +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO Connected all trees +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Connected all rings +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO Connected all trees +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509876:2510279 [7] NCCL INFO comm 0xb92de60 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56724530d017e8f1 - Init COMPLETE +ip-26-0-171-21:1183728:1184124 [1] NCCL INFO comm 0xc265d60 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeb22685fc134aa4a - Init COMPLETE +ip-26-0-171-21:1183734:1184128 [7] NCCL INFO comm 0xa6e78e0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56724530d017e8f1 - Init COMPLETE +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO Connected all trees +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Connected all rings +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO Connected all trees +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Connected all rings +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO Connected all trees +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Connected all rings +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO Connected all trees +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Connected all rings +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO Connected all trees +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Connected all rings +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO Connected all trees +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Connected all rings +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO Connected all trees +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Connected all rings +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO Connected all trees +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Connected all rings +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO Connected all trees +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509869:2510270 [0] NCCL INFO comm 0xba63d80 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf1eadbc30220ebb9 - Init COMPLETE +ip-26-0-168-52:2509875:2510283 [6] NCCL INFO comm 0xb8e7f60 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x866e037d1ddeeb05 - Init COMPLETE +ip-26-0-171-21:1183727:1184125 [0] NCCL INFO comm 0xb6417e0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf1eadbc30220ebb9 - Init COMPLETE +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Connected all rings +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO Connected all trees +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183730:1184131 [3] NCCL INFO comm 0xb75ecb0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfe7d2120cde959e - Init COMPLETE +ip-26-0-171-21:1183731:1184129 [4] NCCL INFO comm 0xa45d980 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4bf8d575fb167966 - Init COMPLETE +ip-26-0-168-52:2509873:2510281 [4] NCCL INFO comm 0xb841de0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4bf8d575fb167966 - Init COMPLETE +ip-26-0-168-52:2509872:2510285 [3] NCCL INFO comm 0xac94270 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfe7d2120cde959e - Init COMPLETE +ip-26-0-171-21:1183733:1184130 [6] NCCL INFO comm 0xa6ffc60 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x866e037d1ddeeb05 - Init COMPLETE +ip-26-0-171-21:1183729:1184127 [2] NCCL INFO comm 0xb778ba0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd8cd3a80175e00f7 - Init COMPLETE +ip-26-0-168-52:2509871:2510277 [2] NCCL INFO comm 0xbf4f860 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xd8cd3a80175e00f7 - Init COMPLETE +12/28/2024 01:07:52 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-21]: Local number of parameters: 604M (1152.04MiB) +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Total number of parameters: 1.61G (3072.13MiB) +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Local number of parameters: 1.01G (1920.09MiB) +12/28/2024 01:07:52 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-21]: [After model building] Memory usage: 1152.06MiB. Peak allocated: 5504.00MiB Peak reserved: 12674.00MiB +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [After model building] Memory usage: 1920.11MiB. Peak allocated: 5504.00MiB Peak reserved: 12674.00MiB +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: No checkpoint path provided. +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Parametrizing model parameters using StandardParametrizator +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [Optimizer Building] Using LearningRateForSP as learning rate +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [ZeRO sharding] Size of optimizer params per rank: +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [ZeRO sharding] DP Rank 0 has 126M out of 1.01G (12.50%) params' optimizer states +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [ZeRO sharding] DP Rank 1 has 126M out of 1.01G (12.50%) params' optimizer states +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [ZeRO sharding] DP Rank 2 has 126M out of 1.01G (12.50%) params' optimizer states +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [ZeRO sharding] DP Rank 3 has 126M out of 1.01G (12.50%) params' optimizer states +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [ZeRO sharding] DP Rank 4 has 126M out of 1.01G (12.50%) params' optimizer states +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [ZeRO sharding] DP Rank 5 has 126M out of 1.01G (12.50%) params' optimizer states +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [ZeRO sharding] DP Rank 6 has 126M out of 1.01G (12.50%) params' optimizer states +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [ZeRO sharding] DP Rank 7 has 126M out of 1.01G (12.50%) params' optimizer states +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.token_position_embeddings | PP: 0/2 | Block rank: 0 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.0 | PP: 0/2 | Block rank: 1 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.1 | PP: 0/2 | Block rank: 2 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.2 | PP: 0/2 | Block rank: 3 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.3 | PP: 0/2 | Block rank: 4 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.4 | PP: 0/2 | Block rank: 5 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.5 | PP: 0/2 | Block rank: 6 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.6 | PP: 0/2 | Block rank: 7 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.7 | PP: 0/2 | Block rank: 8 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.8 | PP: 0/2 | Block rank: 9 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.9 | PP: 0/2 | Block rank: 10 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.10 | PP: 0/2 | Block rank: 11 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.11 | PP: 1/2 | Block rank: 0 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.12 | PP: 1/2 | Block rank: 1 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.13 | PP: 1/2 | Block rank: 2 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.14 | PP: 1/2 | Block rank: 3 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.decoder.15 | PP: 1/2 | Block rank: 4 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.final_layer_norm | PP: 1/2 | Block rank: 5 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.lm_head | PP: 1/2 | Block rank: 6 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: model.cast_to_fp32 | PP: 1/2 | Block rank: 7 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: module_name: loss | PP: 1/2 | Block rank: 8 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Using dummy data generator +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [Training Plan] There are 1 training stages +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [Stage Stable Training Stage] start from step 1 +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: +12/28/2024 01:07:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: [Start training] datetime: 2024-12-28 01:07:52.927740 | mbs: 1 | grad_accum: 32 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/28/2024 01:07:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/28/2024 01:07:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Memory usage: 6240.30MiB. Peak allocated 6240.30MiB. Peak reserved: 16516.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO comm 0xb3354d0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56ae5ea11f353b72 - Init START +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO comm 0x1883ee50 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56ae5ea11f353b72 - Init START +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO comm 0xb092670 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x89acda41776e38b0 - Init START +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO comm 0x1837a120 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x89acda41776e38b0 - Init START +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO comm 0xb31d6c0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xef68ba324b5b3d6d - Init START +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO comm 0x187d4320 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xef68ba324b5b3d6d - Init START +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO comm 0xbde1980 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1d5947c6d7916886 - Init START +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO comm 0x1850bc10 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1d5947c6d7916886 - Init START +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO comm 0xc395820 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbbb3e223727c632 - Init START +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO comm 0x18114060 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbbb3e223727c632 - Init START +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO comm 0xc3b1610 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8636b3cd68e58d2d - Init START +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO comm 0x19200a10 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8636b3cd68e58d2d - Init START +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO comm 0xce9be10 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa8e3a68e0b89eded - Init START +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO comm 0x17b08490 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa8e3a68e0b89eded - Init START +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Using network Libfabric +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO comm 0xc291b60 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa62a4e8f00bb006d - Init START +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO comm 0x193f0360 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa62a4e8f00bb006d - Init START +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/205 +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/184 +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Connected all rings +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO Connected all trees +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Connected all rings +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO Connected all trees +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509875:2510323 [6] NCCL INFO comm 0x1883ee50 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56ae5ea11f353b72 - Init COMPLETE +ip-26-0-168-52:2509875:2510355 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-168-52:2509875:2510355 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-171-21:1183733:1184159 [6] NCCL INFO comm 0xb3354d0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56ae5ea11f353b72 - Init COMPLETE +ip-26-0-171-21:1183733:1184189 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-171-21:1183733:1184189 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Connected all rings +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO Connected all trees +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Connected all rings +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO Connected all trees +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Connected all rings +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO Connected all trees +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Connected all rings +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO Connected all trees +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Connected all rings +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO Connected all trees +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509873:2510325 [4] NCCL INFO comm 0x1837a120 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x89acda41776e38b0 - Init COMPLETE +ip-26-0-168-52:2509873:2510356 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-168-52:2509873:2510356 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-171-21:1183731:1184160 [4] NCCL INFO comm 0xb092670 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x89acda41776e38b0 - Init COMPLETE +ip-26-0-171-21:1183731:1184190 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-171-21:1183731:1184190 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Connected all rings +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO Connected all trees +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509876:2510327 [7] NCCL INFO comm 0x187d4320 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xef68ba324b5b3d6d - Init COMPLETE +ip-26-0-168-52:2509876:2510357 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-168-52:2509876:2510357 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-171-21:1183734:1184162 [7] NCCL INFO comm 0xb31d6c0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xef68ba324b5b3d6d - Init COMPLETE +ip-26-0-171-21:1183734:1184191 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-171-21:1183734:1184191 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-168-52:2509874:2510329 [5] NCCL INFO comm 0x1850bc10 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1d5947c6d7916886 - Init COMPLETE +ip-26-0-168-52:2509874:2510358 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-168-52:2509874:2510358 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-171-21:1183732:1184163 [5] NCCL INFO comm 0xbde1980 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x1d5947c6d7916886 - Init COMPLETE +ip-26-0-171-21:1183732:1184192 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-171-21:1183732:1184192 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Connected all rings +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO Connected all trees +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Connected all rings +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO Connected all trees +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183730:1184168 [3] NCCL INFO comm 0xc395820 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbbb3e223727c632 - Init COMPLETE +ip-26-0-171-21:1183730:1184193 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-171-21:1183730:1184193 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-168-52:2509872:2510333 [3] NCCL INFO comm 0x18114060 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xbbb3e223727c632 - Init COMPLETE +ip-26-0-168-52:2509872:2510359 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-168-52:2509872:2510359 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Connected all rings +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO Connected all trees +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Connected all rings +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO Connected all trees +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183729:1184170 [2] NCCL INFO comm 0xc3b1610 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8636b3cd68e58d2d - Init COMPLETE +ip-26-0-171-21:1183729:1184194 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-171-21:1183729:1184194 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-168-52:2509871:2510336 [2] NCCL INFO comm 0x19200a10 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8636b3cd68e58d2d - Init COMPLETE +ip-26-0-168-52:2509871:2510360 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-168-52:2509871:2510360 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Connected all rings +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO Connected all trees +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Connected all rings +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO Connected all trees +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183728:1184174 [1] NCCL INFO comm 0xce9be10 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa8e3a68e0b89eded - Init COMPLETE +ip-26-0-171-21:1183728:1184195 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-171-21:1183728:1184195 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-168-52:2509870:2510340 [1] NCCL INFO comm 0x17b08490 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa8e3a68e0b89eded - Init COMPLETE +ip-26-0-168-52:2509870:2510361 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-168-52:2509870:2510361 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Connected all rings +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO Connected all trees +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Connected all rings +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO Connected all trees +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-171-21:1183727:1184178 [0] NCCL INFO comm 0xc291b60 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa62a4e8f00bb006d - Init COMPLETE +ip-26-0-168-52:2509869:2510345 [0] NCCL INFO comm 0x193f0360 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa62a4e8f00bb006d - Init COMPLETE +ip-26-0-171-21:1183727:1184196 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-171-21:1183727:1184196 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-168-52:2509869:2510362 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-168-52:2509869:2510362 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-168-52:2509875:2510431 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-168-52:2509875:2510431 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-168-52:2509872:2510432 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-168-52:2509872:2510432 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-168-52:2509873:2510433 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-168-52:2509873:2510433 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-168-52:2509876:2510434 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-168-52:2509876:2510434 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-168-52:2509874:2510435 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-168-52:2509874:2510435 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-168-52:2509871:2510436 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-168-52:2509871:2510436 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-168-52:2509869:2510437 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-168-52:2509869:2510437 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-171-21:1183733:1184281 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-171-21:1183733:1184281 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-171-21:1183731:1184282 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-171-21:1183731:1184282 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-171-21:1183734:1184283 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-171-21:1183734:1184283 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-171-21:1183732:1184284 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-171-21:1183732:1184284 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-168-52:2509870:2510438 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-168-52:2509870:2510438 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-171-21:1183730:1184285 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-171-21:1183730:1184285 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-171-21:1183729:1184286 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-171-21:1183729:1184286 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-171-21:1183727:1184287 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-171-21:1183727:1184287 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-171-21:1183728:1184288 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-171-21:1183728:1184288 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +12/28/2024 01:08:15 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Memory usage: 6309.83MiB. Peak allocated 15649.59MiB. Peak reserved: 16438.00MiB +12/28/2024 01:08:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Memory usage: 7269.91MiB. Peak allocated 11350.09MiB. Peak reserved: 19812.00MiB +12/28/2024 01:08:16 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-21]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 21.6K | tokens_per_sec: 48.5K | tokens_per_sec_per_gpu: 3.03K | global_batch_size: 256 | lm_loss: 12 | lr: 0.00015 | model_tflops_per_gpu: 29.3 | hardware_tflops_per_gpu: 29.3 | grad_norm: 0.847 | cuda_memory_allocated: 4.6G | cuda_max_memory_reserved: 13.7G | hd_total_memory_tb: 312G | hd_used_memory_tb: 190G | hd_free_memory_tb: 122G +12/28/2024 01:08:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Memory usage: 7269.91MiB. Peak allocated 16609.67MiB. Peak reserved: 19812.00MiB +12/28/2024 01:08:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Memory usage: 7269.91MiB. Peak allocated 11350.09MiB. Peak reserved: 19812.00MiB +12/28/2024 01:08:34 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-21]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 18.4K | tokens_per_sec: 56.8K | tokens_per_sec_per_gpu: 3.55K | global_batch_size: 256 | lm_loss: 12 | lr: 0.0003 | model_tflops_per_gpu: 34.3 | hardware_tflops_per_gpu: 34.3 | grad_norm: 0.847 | cuda_memory_allocated: 4.6G | cuda_max_memory_reserved: 13.7G | hd_total_memory_tb: 312G | hd_used_memory_tb: 190G | hd_free_memory_tb: 122G +12/28/2024 01:08:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Memory usage: 7269.91MiB. Peak allocated 16609.67MiB. Peak reserved: 19812.00MiB +12/28/2024 01:08:53 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-21]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 18.5K | tokens_per_sec: 56.7K | tokens_per_sec_per_gpu: 3.54K | global_batch_size: 256 | lm_loss: 12 | lr: 0.000296 | model_tflops_per_gpu: 34.2 | hardware_tflops_per_gpu: 34.2 | grad_norm: 0.839 | cuda_memory_allocated: 4.6G | cuda_max_memory_reserved: 13.7G | hd_total_memory_tb: 312G | hd_used_memory_tb: 190G | hd_free_memory_tb: 122G +12/28/2024 01:08:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/28/2024 01:08:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: | -------- | -------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | ----------------------------------- | -------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/28/2024 01:08:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: | 13848317 | 1.34G_dp8_tp1_pp2_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k | 2 | 4096 | 1 | 32 | 256 | 34.24 | 34.24 | 3542.72 | 331.48 | 162.92 | 160.70 | 422.18 | 239.37 | 240.83 | 11.08 | 19.35 | 8 | 2 | 1 | OneForwardOneBackwardPipelineEngine | REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 32 | 4096 | 131072 | True | torch.bfloat16 | 1 | 25 | True | 1.61G | 1.01G | +12/28/2024 01:08:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final2.csv +12/28/2024 01:08:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Acquired lock for CSV file: benchmark/results/bench_final2.csv +12/28/2024 01:08:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Successfully wrote to CSV file: benchmark/results/bench_final2.csv. Releasing lock... +12/28/2024 01:08:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-168-52]: Throughput logging complete +ip-26-0-168-52:2509870:2510184 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-168-52:2509870:2510054 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-168-52:2509870:2510184 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-168-52:2509870:2510054 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +[2024-12-28 01:09:13,313] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1183728 closing signal SIGTERM +[2024-12-28 01:09:13,313] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1183729 closing signal SIGTERM +[2024-12-28 01:09:13,313] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1183730 closing signal SIGTERM +[2024-12-28 01:09:13,313] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1183731 closing signal SIGTERM +[2024-12-28 01:09:13,313] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1183733 closing signal SIGTERM +[2024-12-28 01:09:13,313] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1183734 closing signal SIGTERM +[2024-12-28 01:09:13,320] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2509870 closing signal SIGTERM +[2024-12-28 01:09:13,320] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2509871 closing signal SIGTERM +[2024-12-28 01:09:13,320] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2509872 closing signal SIGTERM +[2024-12-28 01:09:13,320] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2509873 closing signal SIGTERM +[2024-12-28 01:09:13,320] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2509874 closing signal SIGTERM +[2024-12-28 01:09:13,321] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2509875 closing signal SIGTERM +[2024-12-28 01:09:15,257] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1183727) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-28_01:09:13 + host : ip-26-0-171-21.ec2.internal + rank : 13 (local_rank: 5) + exitcode : 1 (pid: 1183732) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-28_01:09:13 + host : ip-26-0-171-21.ec2.internal + rank : 8 (local_rank: 0) + exitcode : 1 (pid: 1183727) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-28 01:09:15,401] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2509869) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-28_01:09:13 + host : ip-26-0-168-52.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 2509876) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-28_01:09:13 + host : ip-26-0-168-52.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 2509869) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-171-21: task 1: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13848317.0 +slurmstepd: error: *** STEP 13848317.0 ON ip-26-0-168-52 CANCELLED AT 2024-12-28T01:09:15 *** +srun: error: ip-26-0-168-52: task 0: Terminated +srun: Force Terminated StepId=13848317.0 diff --git a/logs/13848463-bench_1.34G_dp4_tp1_pp8_acc32_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13848463-bench_1.34G_dp4_tp1_pp8_acc32_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..54ebab44a09df84c61837ef66147ede74133ee88 --- /dev/null +++ b/logs/13848463-bench_1.34G_dp4_tp1_pp8_acc32_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,19312 @@ ++ '[' -z 13848463 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-161-138,ip-26-0-163-[58,220],ip-26-0-169-239' ++ export 'NODELIST=ip-26-0-161-138 +ip-26-0-163-58 +ip-26-0-163-220 +ip-26-0-169-239' ++ NODELIST='ip-26-0-161-138 +ip-26-0-163-58 +ip-26-0-163-220 +ip-26-0-169-239' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-161-138,ip-26-0-163-[58,220],ip-26-0-169-239' ++ export MASTER_NODE=ip-26-0-161-138 ++ MASTER_NODE=ip-26-0-161-138 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-43 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 52-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 52-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 52-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 52-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-161-138' +Master node: ip-26-0-161-138 ++ echo 'All nodes: ip-26-0-161-138 +ip-26-0-163-58 +ip-26-0-163-220 +ip-26-0-169-239' +All nodes: ip-26-0-161-138 +ip-26-0-163-58 +ip-26-0-163-220 +ip-26-0-169-239 ++ echo 'World size: 32' +World size: 32 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13848463 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-161-138:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.34G_dp4_tp1_pp8_acc32_mbs2_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-28 01:46:43,281] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 01:46:43,281] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 01:46:43,281] torch.distributed.run: [WARNING] +[2024-12-28 01:46:43,281] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 01:46:43,281] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 01:46:43,281] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 01:46:43,281] torch.distributed.run: [WARNING] +[2024-12-28 01:46:43,281] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 01:46:43,281] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 01:46:43,281] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 01:46:43,354] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 01:46:43,354] torch.distributed.run: [WARNING] +[2024-12-28 01:46:43,354] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 01:46:43,354] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 01:46:43,354] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 01:46:43,485] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 01:46:43,485] torch.distributed.run: [WARNING] +[2024-12-28 01:46:43,485] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 01:46:43,485] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 01:46:43,485] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-161-138:1967686:1967686 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-161-138:1967686:1967686 [0] NCCL INFO Bootstrap : Using enp74s0:26.0.173.132<0> +ip-26-0-161-138:1967686:1967686 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-138:1967686:1967686 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-138:1967686:1967686 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:940533:940533 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:940533:940533 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-161-138:1967693:1967693 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-138:1967692:1967692 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:940536:940536 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-138:1967693:1967693 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-161-138:1967689:1967689 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:940533:940533 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-161-138:1967692:1967692 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-58:940536:940536 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-161-138:1967689:1967689 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-161-138:1967690:1967690 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:940530:940530 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-138:1967687:1967687 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-138:1967688:1967688 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:940533:940533 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:940533:940533 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-138:1967690:1967690 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-58:940530:940530 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-161-138:1967687:1967687 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-161-138:1967691:1967691 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-138:1967688:1967688 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-58:940536:940536 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-161-138:1967691:1967691 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-58:940530:940530 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-163-58:940536:940536 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:940536:940536 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-58:940530:940530 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:940530:940530 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-138:1967693:1967693 [7] NCCL INFO Bootstrap : Using enp74s0:26.0.173.132<0> +ip-26-0-161-138:1967692:1967692 [6] NCCL INFO Bootstrap : Using enp74s0:26.0.173.132<0> +ip-26-0-161-138:1967693:1967693 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-138:1967693:1967693 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-58:940532:940532 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-138:1967689:1967689 [3] NCCL INFO Bootstrap : Using enp74s0:26.0.173.132<0> +ip-26-0-169-239:2207359:2207359 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-138:1967690:1967690 [4] NCCL INFO Bootstrap : Using enp74s0:26.0.173.132<0> +ip-26-0-161-138:1967687:1967687 [1] NCCL INFO Bootstrap : Using enp74s0:26.0.173.132<0> +ip-26-0-163-58:940535:940535 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-138:1967688:1967688 [2] NCCL INFO Bootstrap : Using enp74s0:26.0.173.132<0> +ip-26-0-161-138:1967691:1967691 [5] NCCL INFO Bootstrap : Using enp74s0:26.0.173.132<0> +ip-26-0-163-58:940532:940532 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-239:2207361:2207361 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-239:2207364:2207364 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-138:1967692:1967692 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-138:1967692:1967692 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-58:940529:940529 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:940535:940535 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-239:2207363:2207363 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-138:1967689:1967689 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-138:1967689:1967689 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-58:940529:940529 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-239:2207359:2207359 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-239:2207364:2207364 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-220:1685421:1685421 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-239:2207363:2207363 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-239:2207361:2207361 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-161-138:1967687:1967687 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-138:1967687:1967687 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1685420:1685420 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-161-138:1967690:1967690 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-138:1967688:1967688 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-138:1967691:1967691 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-161-138:1967690:1967690 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-138:1967688:1967688 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-138:1967691:1967691 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-58:940534:940534 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:940532:940532 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-163-220:1685418:1685418 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1685416:1685416 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-239:2207362:2207362 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1685421:1685421 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-220:1685419:1685419 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1685414:1685414 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:940531:940531 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1685417:1685417 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-239:2207357:2207357 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1685415:1685415 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-220:1685420:1685420 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-58:940534:940534 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-220:1685416:1685416 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-220:1685418:1685418 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-58:940535:940535 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-169-239:2207362:2207362 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-58:940532:940532 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:940532:940532 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1685419:1685419 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-220:1685414:1685414 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-58:940531:940531 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-220:1685415:1685415 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-239:2207357:2207357 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-220:1685417:1685417 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-58:940535:940535 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:940535:940535 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-239:2207358:2207358 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-163-58:940529:940529 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-169-239:2207360:2207360 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-169-239:2207358:2207358 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-169-239:2207360:2207360 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-163-58:940529:940529 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:940529:940529 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-58:940534:940534 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-163-58:940531:940531 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.163.58<0> +ip-26-0-163-58:940534:940534 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:940534:940534 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-58:940531:940531 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-58:940531:940531 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-239:2207364:2207364 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.169.239<0> +ip-26-0-169-239:2207359:2207359 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.169.239<0> +ip-26-0-169-239:2207363:2207363 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.169.239<0> +ip-26-0-169-239:2207361:2207361 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.169.239<0> +ip-26-0-169-239:2207359:2207359 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-239:2207364:2207364 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-239:2207359:2207359 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-239:2207364:2207364 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-239:2207361:2207361 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-239:2207363:2207363 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-239:2207363:2207363 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-239:2207361:2207361 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-239:2207362:2207362 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.169.239<0> +ip-26-0-169-239:2207357:2207357 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.169.239<0> +ip-26-0-169-239:2207358:2207358 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.169.239<0> +ip-26-0-169-239:2207360:2207360 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.169.239<0> +ip-26-0-163-220:1685421:1685421 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-169-239:2207358:2207358 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-239:2207362:2207362 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-239:2207362:2207362 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-239:2207358:2207358 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-239:2207357:2207357 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-239:2207360:2207360 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-169-239:2207357:2207357 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-169-239:2207360:2207360 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1685420:1685420 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-220:1685421:1685421 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1685421:1685421 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1685416:1685416 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-220:1685418:1685418 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-220:1685414:1685414 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-220:1685415:1685415 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-220:1685417:1685417 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-220:1685419:1685419 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.163.220<0> +ip-26-0-163-220:1685420:1685420 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1685420:1685420 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1685416:1685416 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1685418:1685418 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1685418:1685418 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1685416:1685416 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1685415:1685415 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1685415:1685415 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1685419:1685419 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1685414:1685414 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1685419:1685419 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1685414:1685414 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-163-220:1685417:1685417 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-163-220:1685417:1685417 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/253 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:940536:940688 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:940536:940688 [7] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Using network Libfabric +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:940530:940690 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:940530:940690 [1] NCCL INFO Using network Libfabric +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:940532:940691 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:940532:940691 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:940529:940692 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:940529:940692 [0] NCCL INFO Using network Libfabric +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:940534:940693 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:940534:940693 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:940533:940687 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:940533:940687 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:940531:940694 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:940531:940694 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-58:940535:940689 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-58:940535:940689 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO comm 0x8ed3010 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO comm 0x9b395e0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO comm 0xa2e8e60 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-58:940529:940692 [0] NCCL INFO comm 0x871acf0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-58:940530:940690 [1] NCCL INFO comm 0x8d62e30 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-58:940532:940691 [3] NCCL INFO comm 0x8757f10 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-58:940531:940694 [2] NCCL INFO comm 0x9a08f20 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-58:940533:940687 [4] NCCL INFO comm 0x88252a0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-58:940535:940689 [6] NCCL INFO comm 0xa188420 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-58:940534:940693 [5] NCCL INFO comm 0xa3c5b00 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-58:940536:940688 [7] NCCL INFO comm 0x8952a40 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x18a55e82202f487c - Init START +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO comm 0xa3be000 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x18a55e82202f487c - Init START +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO comm 0xa3798c0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x18a55e82202f487c - Init START +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO comm 0x98a2010 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x18a55e82202f487c - Init START +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO comm 0x9010950 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x18a55e82202f487c - Init START +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO comm 0x96e5bc0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x18a55e82202f487c - Init START +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO comm 0x8595290 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x18a55e82202f487c - Init START +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO comm 0x964afc0 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO comm 0x97a1620 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO comm 0x97ee680 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x18a55e82202f487c - Init START +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO comm 0x88e1230 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x18a55e82202f487c - Init START +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO comm 0xa209610 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x18a55e82202f487c - Init START +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO comm 0xa04ae00 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x18a55e82202f487c - Init START +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO comm 0xa3427f0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x18a55e82202f487c - Init START +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO comm 0xa154170 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x18a55e82202f487c - Init START +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO comm 0xa143c30 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x18a55e82202f487c - Init START +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO comm 0x90fbf10 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x18a55e82202f487c - Init START +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO comm 0x950fba0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x18a55e82202f487c - Init START +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO comm 0x9c5b1b0 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO comm 0xa0de160 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO comm 0x85fa070 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO comm 0x955ee10 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x18a55e82202f487c - Init START +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940532:940691 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940531:940694 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940529:940692 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-163-58:940532:940691 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:940532:940691 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-163-58:940533:940687 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940533:940687 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-163-58:940535:940689 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940535:940689 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-163-58:940536:940688 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940536:940688 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-163-58:940530:940690 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:940534:940693 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940534:940693 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-163-58:940530:940690 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-163-58:940529:940692 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-58:940529:940692 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-163-58:940531:940694 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-58:940531:940694 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-163-58:940529:940692 [0] NCCL INFO Trees [0] 17/8/24->16->0 [1] -1/-1/-1->16->23 [2] 17/-1/-1->16->23 [3] 17/-1/-1->16->23 [4] 17/-1/-1->16->23 [5] 17/-1/-1->16->23 [6] 17/-1/-1->16->23 [7] 17/-1/-1->16->23 [8] 17/-1/-1->16->8 [9] -1/-1/-1->16->23 [10] 17/-1/-1->16->23 [11] 17/-1/-1->16->23 [12] 17/-1/-1->16->23 [13] 17/-1/-1->16->23 [14] 17/-1/-1->16->23 [15] 17/-1/-1->16->23 +ip-26-0-163-58:940529:940692 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940530:940690 [1] NCCL INFO Trees [0] 18/-1/-1->17->16 [1] 18/9/25->17->1 [2] -1/-1/-1->17->16 [3] 18/-1/-1->17->16 [4] 18/-1/-1->17->16 [5] 18/-1/-1->17->16 [6] 18/-1/-1->17->16 [7] 18/-1/-1->17->16 [8] 18/-1/-1->17->16 [9] 18/-1/-1->17->9 [10] -1/-1/-1->17->16 [11] 18/-1/-1->17->16 [12] 18/-1/-1->17->16 [13] 18/-1/-1->17->16 [14] 18/-1/-1->17->16 [15] 18/-1/-1->17->16 +ip-26-0-163-58:940530:940690 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->22 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/22/6->14->30 [15] -1/-1/-1->14->13 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Trees [0] 26/-1/-1->25->24 [1] 26/-1/-1->25->17 [2] -1/-1/-1->25->24 [3] 26/-1/-1->25->24 [4] 26/-1/-1->25->24 [5] 26/-1/-1->25->24 [6] 26/-1/-1->25->24 [7] 26/-1/-1->25->24 [8] 26/-1/-1->25->24 [9] 26/9/-1->25->-1 [10] -1/-1/-1->25->24 [11] 26/-1/-1->25->24 [12] 26/-1/-1->25->24 [13] 26/-1/-1->25->24 [14] 26/-1/-1->25->24 [15] 26/-1/-1->25->24 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Trees [0] 25/-1/-1->24->16 [1] -1/-1/-1->24->31 [2] 25/-1/-1->24->31 [3] 25/-1/-1->24->31 [4] 25/-1/-1->24->31 [5] 25/-1/-1->24->31 [6] 25/-1/-1->24->31 [7] 25/-1/-1->24->31 [8] 25/8/-1->24->-1 [9] -1/-1/-1->24->31 [10] 25/-1/-1->24->31 [11] 25/-1/-1->24->31 [12] 25/-1/-1->24->31 [13] 25/-1/-1->24->31 [14] 25/-1/-1->24->31 [15] 25/-1/-1->24->31 +ip-26-0-163-58:940531:940694 [2] NCCL INFO Trees [0] 19/-1/-1->18->17 [1] 19/-1/-1->18->17 [2] 19/10/26->18->2 [3] -1/-1/-1->18->17 [4] 19/-1/-1->18->17 [5] 19/-1/-1->18->17 [6] 19/-1/-1->18->17 [7] 19/-1/-1->18->17 [8] 19/-1/-1->18->17 [9] 19/-1/-1->18->17 [10] 19/-1/-1->18->10 [11] -1/-1/-1->18->17 [12] 19/-1/-1->18->17 [13] 19/-1/-1->18->17 [14] 19/-1/-1->18->17 [15] 19/-1/-1->18->17 +ip-26-0-163-58:940531:940694 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/23/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 +ip-26-0-163-58:940536:940688 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] 16/-1/-1->23->22 [2] 16/-1/-1->23->22 [3] 16/-1/-1->23->22 [4] 16/-1/-1->23->22 [5] 16/-1/-1->23->22 [6] 16/-1/-1->23->22 [7] 16/15/31->23->7 [8] -1/-1/-1->23->22 [9] 16/-1/-1->23->22 [10] 16/-1/-1->23->22 [11] 16/-1/-1->23->22 [12] 16/-1/-1->23->22 [13] 16/-1/-1->23->22 [14] 16/-1/-1->23->22 [15] 16/-1/-1->23->15 +ip-26-0-163-58:940536:940688 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940532:940691 [3] NCCL INFO Trees [0] 20/-1/-1->19->18 [1] 20/-1/-1->19->18 [2] 20/-1/-1->19->18 [3] 20/11/27->19->3 [4] -1/-1/-1->19->18 [5] 20/-1/-1->19->18 [6] 20/-1/-1->19->18 [7] 20/-1/-1->19->18 [8] 20/-1/-1->19->18 [9] 20/-1/-1->19->18 [10] 20/-1/-1->19->18 [11] 20/-1/-1->19->11 [12] -1/-1/-1->19->18 [13] 20/-1/-1->19->18 [14] 20/-1/-1->19->18 [15] 20/-1/-1->19->18 +ip-26-0-163-58:940532:940691 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Trees [0] 27/-1/-1->26->25 [1] 27/-1/-1->26->25 [2] 27/-1/-1->26->18 [3] -1/-1/-1->26->25 [4] 27/-1/-1->26->25 [5] 27/-1/-1->26->25 [6] 27/-1/-1->26->25 [7] 27/-1/-1->26->25 [8] 27/-1/-1->26->25 [9] 27/-1/-1->26->25 [10] 27/10/-1->26->-1 [11] -1/-1/-1->26->25 [12] 27/-1/-1->26->25 [13] 27/-1/-1->26->25 [14] 27/-1/-1->26->25 [15] 27/-1/-1->26->25 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940534:940693 [5] NCCL INFO Trees [0] 22/-1/-1->21->20 [1] 22/-1/-1->21->20 [2] 22/-1/-1->21->20 [3] 22/-1/-1->21->20 [4] 22/-1/-1->21->20 [5] 22/13/29->21->5 [6] -1/-1/-1->21->20 [7] 22/-1/-1->21->20 [8] 22/-1/-1->21->20 [9] 22/-1/-1->21->20 [10] 22/-1/-1->21->20 [11] 22/-1/-1->21->20 [12] 22/-1/-1->21->20 [13] 22/-1/-1->21->13 [14] -1/-1/-1->21->20 [15] 22/-1/-1->21->20 +ip-26-0-163-58:940533:940687 [4] NCCL INFO Trees [0] 21/-1/-1->20->19 [1] 21/-1/-1->20->19 [2] 21/-1/-1->20->19 [3] 21/-1/-1->20->19 [4] 21/12/28->20->4 [5] -1/-1/-1->20->19 [6] 21/-1/-1->20->19 [7] 21/-1/-1->20->19 [8] 21/-1/-1->20->19 [9] 21/-1/-1->20->19 [10] 21/-1/-1->20->19 [11] 21/-1/-1->20->19 [12] 21/-1/-1->20->12 [13] -1/-1/-1->20->19 [14] 21/-1/-1->20->19 [15] 21/-1/-1->20->19 +ip-26-0-163-58:940534:940693 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940533:940687 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->23 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/23/7->15->31 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->18 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/18/2->10->26 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] 24/-1/-1->31->30 [2] 24/-1/-1->31->30 [3] 24/-1/-1->31->30 [4] 24/-1/-1->31->30 [5] 24/-1/-1->31->30 [6] 24/-1/-1->31->30 [7] 24/-1/-1->31->23 [8] -1/-1/-1->31->30 [9] 24/-1/-1->31->30 [10] 24/-1/-1->31->30 [11] 24/-1/-1->31->30 [12] 24/-1/-1->31->30 [13] 24/-1/-1->31->30 [14] 24/-1/-1->31->30 [15] 24/15/-1->31->-1 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Trees [0] 31/-1/-1->30->29 [1] 31/-1/-1->30->29 [2] 31/-1/-1->30->29 [3] 31/-1/-1->30->29 [4] 31/-1/-1->30->29 [5] 31/-1/-1->30->29 [6] 31/-1/-1->30->22 [7] -1/-1/-1->30->29 [8] 31/-1/-1->30->29 [9] 31/-1/-1->30->29 [10] 31/-1/-1->30->29 [11] 31/-1/-1->30->29 [12] 31/-1/-1->30->29 [13] 31/-1/-1->30->29 [14] 31/14/-1->30->-1 [15] -1/-1/-1->30->29 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940535:940689 [6] NCCL INFO Trees [0] 23/-1/-1->22->21 [1] 23/-1/-1->22->21 [2] 23/-1/-1->22->21 [3] 23/-1/-1->22->21 [4] 23/-1/-1->22->21 [5] 23/-1/-1->22->21 [6] 23/14/30->22->6 [7] -1/-1/-1->22->21 [8] 23/-1/-1->22->21 [9] 23/-1/-1->22->21 [10] 23/-1/-1->22->21 [11] 23/-1/-1->22->21 [12] 23/-1/-1->22->21 [13] 23/-1/-1->22->21 [14] 23/-1/-1->22->14 [15] -1/-1/-1->22->21 +ip-26-0-163-58:940535:940689 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->20 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/20/4->12->28 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->17 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/17/1->9->25 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Trees [0] 29/-1/-1->28->27 [1] 29/-1/-1->28->27 [2] 29/-1/-1->28->27 [3] 29/-1/-1->28->27 [4] 29/-1/-1->28->20 [5] -1/-1/-1->28->27 [6] 29/-1/-1->28->27 [7] 29/-1/-1->28->27 [8] 29/-1/-1->28->27 [9] 29/-1/-1->28->27 [10] 29/-1/-1->28->27 [11] 29/-1/-1->28->27 [12] 29/12/-1->28->-1 [13] -1/-1/-1->28->27 [14] 29/-1/-1->28->27 [15] 29/-1/-1->28->27 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Trees [0] 30/-1/-1->29->28 [1] 30/-1/-1->29->28 [2] 30/-1/-1->29->28 [3] 30/-1/-1->29->28 [4] 30/-1/-1->29->28 [5] 30/-1/-1->29->21 [6] -1/-1/-1->29->28 [7] 30/-1/-1->29->28 [8] 30/-1/-1->29->28 [9] 30/-1/-1->29->28 [10] 30/-1/-1->29->28 [11] 30/-1/-1->29->28 [12] 30/-1/-1->29->28 [13] 30/13/-1->29->-1 [14] -1/-1/-1->29->28 [15] 30/-1/-1->29->28 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->19 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/19/3->11->27 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->21 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/21/5->13->29 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/22/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Trees [0] 28/-1/-1->27->26 [1] 28/-1/-1->27->26 [2] 28/-1/-1->27->26 [3] 28/-1/-1->27->19 [4] -1/-1/-1->27->26 [5] 28/-1/-1->27->26 [6] 28/-1/-1->27->26 [7] 28/-1/-1->27->26 [8] 28/-1/-1->27->26 [9] 28/-1/-1->27->26 [10] 28/-1/-1->27->26 [11] 28/11/-1->27->-1 [12] -1/-1/-1->27->26 [13] 28/-1/-1->27->26 [14] 28/-1/-1->27->26 [15] 28/-1/-1->27->26 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Trees [0] 9/-1/-1->8->16 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/16/0->8->24 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 01/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/18/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 03/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/20/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 04/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 05/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/21/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 06/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/19/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 07/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/17/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 09/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 11/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 12/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 13/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 14/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 15/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Trees [0] 1/16/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 08/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 08/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 05/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 08/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 06/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 08/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 13/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 08/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 14/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 05/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 08/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 06/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 05/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 13/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 14/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 06/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 13/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 14/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 08/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 08/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 08/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 08/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 08/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 08/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 05/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:940532:940691 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 06/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 13/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 14/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 08/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 05/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 13/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 05/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 13/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 08/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 05/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 13/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 06/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 06/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 14/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 14/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 07/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 07/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 15/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 06/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 15/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 06/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 14/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 14/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 07/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 07/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 15/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 15/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 01/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 01/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 02/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 03/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 02/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 04/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 03/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 09/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 04/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 09/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 10/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 11/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 12/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 10/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 11/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 12/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 01/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 10/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 04/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 05/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 06/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 07/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 09/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 10/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 01/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 11/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 05/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 12/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 06/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 13/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 04/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 13/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 14/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 05/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 06/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 06/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 05/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 06/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 14/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 06/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 14/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 07/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 14/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 05/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 15/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 13/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 05/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 09/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 14/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 04/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 13/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 05/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 10/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 06/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 13/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 11/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 07/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 09/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 10/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 12/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 11/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 13/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 04/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 12/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 14/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 05/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 13/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 15/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 06/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 14/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 07/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 10/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 11/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 12/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 13/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 14/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 15/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 15/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 07/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 15/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 07/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 07/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 15/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 15/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 02/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 03/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 04/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 02/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 05/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 09/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 10/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 11/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 12/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 13/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 03/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 01/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 03/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 02/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 04/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 04/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 09/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 05/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 10/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 06/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 04/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 04/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 12/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 07/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 07/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 01/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 09/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 09/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 10/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 11/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 11/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 12/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 06/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 15/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 12/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 07/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 13/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 09/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 05/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 09/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 10/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 11/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 14/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 10/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 12/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 13/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 01/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 02/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 04/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 09/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 10/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 12/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 15/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 11/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 03/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 04/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 14/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 05/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 04/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 06/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 07/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 07/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 09/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 09/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 10/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 01/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 11/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 12/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 11/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 12/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 12/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 13/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 15/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 06/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 14/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 07/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 15/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 11/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 09/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 10/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 11/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 14/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 15/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 15/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Connected all rings +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Connected all rings +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Connected all rings +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Connected all rings +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 04/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 05/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 06/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 07/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 10/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 11/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 12/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 13/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 14/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 15/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Connected all rings +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 04/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 05/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Connected all rings +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 06/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Connected all rings +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 04/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 07/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 10/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Connected all rings +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 05/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 06/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 11/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 12/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 13/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 07/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 14/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 15/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 09/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 11/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 12/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 13/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 14/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 15/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 05/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Connected all rings +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Connected all rings +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 06/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 04/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Connected all rings +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 07/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 04/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 05/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Connected all rings +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 05/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 06/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 13/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 06/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 07/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 05/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 14/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 07/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 06/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 15/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 09/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 07/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 09/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 10/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 11/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Connected all rings +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 09/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Connected all rings +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 12/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 12/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 13/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 14/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 13/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 04/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 15/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 05/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 05/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 10/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 06/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 14/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 11/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 06/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 09/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 15/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 13/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 07/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 10/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 14/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 13/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 11/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 15/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 14/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 12/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 13/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 15/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 04/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 14/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 05/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 04/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 04/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 07/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 05/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 06/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 04/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 09/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 07/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 05/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 10/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 06/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 09/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 06/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 11/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 10/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 07/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 09/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 12/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 11/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 09/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 10/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 04/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 13/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 12/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 11/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 15/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Connected all rings +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 10/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 14/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 06/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 12/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 15/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 13/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 14/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 12/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 07/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 13/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 09/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 10/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 14/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 11/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 15/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 12/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 14/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 15/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 04/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 05/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 07/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 09/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 10/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 11/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 12/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 13/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 15/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Connected all rings +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Connected all rings +ip-26-0-163-58:940532:940691 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 05/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 06/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 07/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 09/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 10/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 11/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 13/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 14/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 15/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 00/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 08/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 00/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 08/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Connected all rings +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 08/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 01/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 09/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 01/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 02/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 03/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 08/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 02/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 04/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 09/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 03/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 04/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 05/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 05/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 06/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 05/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 06/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 07/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 08/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 07/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 09/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 13/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 09/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 10/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 05/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 11/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 10/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 12/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 08/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 11/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 13/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 13/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 12/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 13/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 14/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 14/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 15/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 15/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 00/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 00/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 04/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 03/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 08/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 05/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 00/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 06/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 00/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 12/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 06/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 04/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 07/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 08/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 03/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 06/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 08/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 08/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 08/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 11/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 14/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 05/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 12/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 14/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 02/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 06/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 13/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 08/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 02/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 07/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 14/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 08/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 10/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 15/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 08/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 11/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 13/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 10/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 14/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 15/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Connected all trees +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NVLS comm 0xa3798c0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Connected all trees +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NVLS comm 0xa3be000 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Connected all trees +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NVLS comm 0xa209610 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Connected all trees +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NVLS comm 0xa143c30 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Connected all trees +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Connected all trees +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NVLS comm 0x8595290 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Connected all trees +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NVLS comm 0x98a2010 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NVLS comm 0x9010950 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Connected all trees +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NVLS comm 0x96e5bc0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Connected all trees +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NVLS comm 0x90fbf10 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NVLS comm 0x950fba0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NVLS comm 0xa3427f0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Connected all trees +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NVLS comm 0x9c5b1b0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Connected all trees +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NVLS comm 0xa154170 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940530:940690 [1] NCCL INFO Connected all trees +ip-26-0-163-58:940530:940690 [1] NCCL INFO NVLS comm 0x8d62e30 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940536:940688 [7] NCCL INFO Connected all trees +ip-26-0-163-58:940536:940688 [7] NCCL INFO NVLS comm 0x8952a40 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Connected all trees +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NVLS comm 0x964afc0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940535:940689 [6] NCCL INFO Connected all trees +ip-26-0-163-58:940535:940689 [6] NCCL INFO NVLS comm 0xa188420 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940529:940692 [0] NCCL INFO Connected all trees +ip-26-0-163-58:940529:940692 [0] NCCL INFO NVLS comm 0x871acf0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Connected all trees +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NVLS comm 0x88e1230 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Connected all trees +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NVLS comm 0xa04ae00 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940533:940687 [4] NCCL INFO Connected all trees +ip-26-0-163-58:940533:940687 [4] NCCL INFO NVLS comm 0x88252a0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940531:940694 [2] NCCL INFO Connected all trees +ip-26-0-163-58:940531:940694 [2] NCCL INFO NVLS comm 0x9a08f20 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940532:940691 [3] NCCL INFO Connected all trees +ip-26-0-163-58:940532:940691 [3] NCCL INFO NVLS comm 0x8757f10 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NVLS comm 0x955ee10 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NVLS comm 0x9b395e0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NVLS comm 0xa2e8e60 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NVLS comm 0xa0de160 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NVLS comm 0x97a1620 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NVLS comm 0x97ee680 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NVLS comm 0x85fa070 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940534:940693 [5] NCCL INFO Connected all trees +ip-26-0-163-58:940534:940693 [5] NCCL INFO NVLS comm 0xa3c5b00 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NVLS comm 0x8ed3010 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940693 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940530:940690 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940688 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940692 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940691 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940694 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940687 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940689 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940529:940692 [0] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940529:940692 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940529:940692 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940529:940692 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967686:1967851 [0] NCCL INFO comm 0x88e1230 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-161-138:1967690:1967855 [4] NCCL INFO comm 0xa04ae00 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-161-138:1967692:1967857 [6] NCCL INFO comm 0xa209610 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-161-138:1967689:1967853 [3] NCCL INFO comm 0x950fba0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-161-138:1967688:1967854 [2] NCCL INFO comm 0xa143c30 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-161-138:1967691:1967858 [5] NCCL INFO comm 0xa3427f0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-161-138:1967693:1967852 [7] NCCL INFO comm 0xa154170 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-161-138:1967687:1967856 [1] NCCL INFO comm 0x9c5b1b0 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940533:940687 [4] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940533:940687 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940533:940687 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940533:940687 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940531:940694 [2] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940531:940694 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940531:940694 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940531:940694 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940535:940689 [6] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940535:940689 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940535:940689 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940535:940689 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940536:940688 [7] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940536:940688 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940536:940688 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940536:940688 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207364:2207524 [7] NCCL INFO comm 0x964afc0 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-169-239:2207360:2207522 [3] NCCL INFO comm 0x96e5bc0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-169-239:2207362:2207519 [5] NCCL INFO comm 0x90fbf10 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-169-239:2207363:2207518 [6] NCCL INFO comm 0x8595290 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-169-239:2207361:2207517 [4] NCCL INFO comm 0x9010950 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-169-239:2207357:2207521 [0] NCCL INFO comm 0xa3be000 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-169-239:2207359:2207520 [2] NCCL INFO comm 0x98a2010 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-169-239:2207358:2207523 [1] NCCL INFO comm 0xa3798c0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-58:940534:940693 [5] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940534:940693 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940534:940693 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940534:940693 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940530:940690 [1] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940530:940690 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940530:940690 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940530:940690 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940532:940691 [3] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940532:940691 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940532:940691 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940532:940691 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685418:1685581 [4] NCCL INFO comm 0x97a1620 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-220:1685416:1685582 [2] NCCL INFO comm 0x97ee680 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-220:1685420:1685583 [6] NCCL INFO comm 0x9b395e0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-220:1685421:1685579 [7] NCCL INFO comm 0xa2e8e60 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-220:1685417:1685585 [3] NCCL INFO comm 0x85fa070 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-220:1685414:1685580 [0] NCCL INFO comm 0xa0de160 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-220:1685415:1685586 [1] NCCL INFO comm 0x955ee10 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-220:1685419:1685584 [5] NCCL INFO comm 0x8ed3010 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-58:940530:940690 [1] NCCL INFO comm 0x8d62e30 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-58:940534:940693 [5] NCCL INFO comm 0xa3c5b00 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-58:940532:940691 [3] NCCL INFO comm 0x8757f10 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-58:940536:940688 [7] NCCL INFO comm 0x8952a40 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-58:940535:940689 [6] NCCL INFO comm 0xa188420 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-58:940533:940687 [4] NCCL INFO comm 0x88252a0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-58:940529:940692 [0] NCCL INFO comm 0x871acf0 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-163-58:940531:940694 [2] NCCL INFO comm 0x9a08f20 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0x18a55e82202f487c - Init COMPLETE +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:940534:940860 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:940529:940858 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:940532:940861 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:940530:940859 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Using network Libfabric +ip-26-0-163-58:940531:940862 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:940536:940863 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:940533:940865 [4] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Using network Libfabric +ip-26-0-163-58:940535:940864 [6] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO comm 0xa51fec0 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xea96d08e83e9950d - Init START +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO comm 0xa386880 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xea96d08e83e9950d - Init START +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO comm 0xa43e800 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO comm 0x9d70da0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO comm 0x910a920 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO comm 0x99d85c0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO comm 0x88318a0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xea96d08e83e9950d - Init START +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO comm 0xa577ac0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO comm 0x9a261a0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO comm 0x9796520 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO comm 0xa312fb0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xea96d08e83e9950d - Init START +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO comm 0x9744c20 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xea96d08e83e9950d - Init START +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO comm 0xa27e270 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xea96d08e83e9950d - Init START +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO comm 0xa378ac0 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xea96d08e83e9950d - Init START +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO comm 0x9e90ba0 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xea96d08e83e9950d - Init START +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO comm 0x8b20d00 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xea96d08e83e9950d - Init START +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO comm 0x9881ce0 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xea96d08e83e9950d - Init START +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO comm 0x87c9d20 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xea96d08e83e9950d - Init START +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO comm 0x9246d00 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xea96d08e83e9950d - Init START +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO comm 0x932f470 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xea96d08e83e9950d - Init START +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO comm 0x9918980 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xea96d08e83e9950d - Init START +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO comm 0x9ad7240 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-58:940536:940863 [7] NCCL INFO comm 0x8b898b0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xea96d08e83e9950d - Init START +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO comm 0xa5af2f0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xea96d08e83e9950d - Init START +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO comm 0xa5f3ea0 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-58:940535:940864 [6] NCCL INFO comm 0xa3bdc70 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-58:940534:940860 [5] NCCL INFO comm 0xa5fab60 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-58:940533:940865 [4] NCCL INFO comm 0x8a5db40 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-58:940532:940861 [3] NCCL INFO comm 0x898fc20 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-58:940531:940862 [2] NCCL INFO comm 0x9c3e4f0 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-58:940530:940859 [1] NCCL INFO comm 0x8f99900 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-58:940529:940858 [0] NCCL INFO comm 0x8950e00 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xea96d08e83e9950d - Init START +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940536:940863 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940533:940865 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940535:940864 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940534:940860 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940532:940861 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940531:940862 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940529:940858 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940530:940859 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-163-58:940535:940864 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940535:940864 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-163-58:940533:940865 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940533:940865 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-163-58:940532:940861 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:940532:940861 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-163-58:940534:940860 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940534:940860 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-163-58:940530:940859 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:940530:940859 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-163-58:940529:940858 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-58:940529:940858 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-163-58:940531:940862 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-58:940531:940862 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-163-58:940536:940863 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940536:940863 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->23 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/23/7->15->31 +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->22 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/22/6->14->30 [15] -1/-1/-1->14->13 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] 24/-1/-1->31->30 [2] 24/-1/-1->31->30 [3] 24/-1/-1->31->30 [4] 24/-1/-1->31->30 [5] 24/-1/-1->31->30 [6] 24/-1/-1->31->30 [7] 24/-1/-1->31->23 [8] -1/-1/-1->31->30 [9] 24/-1/-1->31->30 [10] 24/-1/-1->31->30 [11] 24/-1/-1->31->30 [12] 24/-1/-1->31->30 [13] 24/-1/-1->31->30 [14] 24/-1/-1->31->30 [15] 24/15/-1->31->-1 +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940536:940863 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] 16/-1/-1->23->22 [2] 16/-1/-1->23->22 [3] 16/-1/-1->23->22 [4] 16/-1/-1->23->22 [5] 16/-1/-1->23->22 [6] 16/-1/-1->23->22 [7] 16/15/31->23->7 [8] -1/-1/-1->23->22 [9] 16/-1/-1->23->22 [10] 16/-1/-1->23->22 [11] 16/-1/-1->23->22 [12] 16/-1/-1->23->22 [13] 16/-1/-1->23->22 [14] 16/-1/-1->23->22 [15] 16/-1/-1->23->15 +ip-26-0-163-58:940533:940865 [4] NCCL INFO Trees [0] 21/-1/-1->20->19 [1] 21/-1/-1->20->19 [2] 21/-1/-1->20->19 [3] 21/-1/-1->20->19 [4] 21/12/28->20->4 [5] -1/-1/-1->20->19 [6] 21/-1/-1->20->19 [7] 21/-1/-1->20->19 [8] 21/-1/-1->20->19 [9] 21/-1/-1->20->19 [10] 21/-1/-1->20->19 [11] 21/-1/-1->20->19 [12] 21/-1/-1->20->12 [13] -1/-1/-1->20->19 [14] 21/-1/-1->20->19 [15] 21/-1/-1->20->19 +ip-26-0-163-58:940536:940863 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940533:940865 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->19 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/19/3->11->27 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->21 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/21/5->13->29 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/23/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/22/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940535:940864 [6] NCCL INFO Trees [0] 23/-1/-1->22->21 [1] 23/-1/-1->22->21 [2] 23/-1/-1->22->21 [3] 23/-1/-1->22->21 [4] 23/-1/-1->22->21 [5] 23/-1/-1->22->21 [6] 23/14/30->22->6 [7] -1/-1/-1->22->21 [8] 23/-1/-1->22->21 [9] 23/-1/-1->22->21 [10] 23/-1/-1->22->21 [11] 23/-1/-1->22->21 [12] 23/-1/-1->22->21 [13] 23/-1/-1->22->21 [14] 23/-1/-1->22->14 [15] -1/-1/-1->22->21 +ip-26-0-163-58:940535:940864 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/20/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Trees [0] 31/-1/-1->30->29 [1] 31/-1/-1->30->29 [2] 31/-1/-1->30->29 [3] 31/-1/-1->30->29 [4] 31/-1/-1->30->29 [5] 31/-1/-1->30->29 [6] 31/-1/-1->30->22 [7] -1/-1/-1->30->29 [8] 31/-1/-1->30->29 [9] 31/-1/-1->30->29 [10] 31/-1/-1->30->29 [11] 31/-1/-1->30->29 [12] 31/-1/-1->30->29 [13] 31/-1/-1->30->29 [14] 31/14/-1->30->-1 [15] -1/-1/-1->30->29 +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940530:940859 [1] NCCL INFO Trees [0] 18/-1/-1->17->16 [1] 18/9/25->17->1 [2] -1/-1/-1->17->16 [3] 18/-1/-1->17->16 [4] 18/-1/-1->17->16 [5] 18/-1/-1->17->16 [6] 18/-1/-1->17->16 [7] 18/-1/-1->17->16 [8] 18/-1/-1->17->16 [9] 18/-1/-1->17->9 [10] -1/-1/-1->17->16 [11] 18/-1/-1->17->16 [12] 18/-1/-1->17->16 [13] 18/-1/-1->17->16 [14] 18/-1/-1->17->16 [15] 18/-1/-1->17->16 +ip-26-0-163-58:940532:940861 [3] NCCL INFO Trees [0] 20/-1/-1->19->18 [1] 20/-1/-1->19->18 [2] 20/-1/-1->19->18 [3] 20/11/27->19->3 [4] -1/-1/-1->19->18 [5] 20/-1/-1->19->18 [6] 20/-1/-1->19->18 [7] 20/-1/-1->19->18 [8] 20/-1/-1->19->18 [9] 20/-1/-1->19->18 [10] 20/-1/-1->19->18 [11] 20/-1/-1->19->11 [12] -1/-1/-1->19->18 [13] 20/-1/-1->19->18 [14] 20/-1/-1->19->18 [15] 20/-1/-1->19->18 +ip-26-0-163-58:940530:940859 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940532:940861 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->17 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/17/1->9->25 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->20 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/20/4->12->28 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/21/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Trees [0] 26/-1/-1->25->24 [1] 26/-1/-1->25->17 [2] -1/-1/-1->25->24 [3] 26/-1/-1->25->24 [4] 26/-1/-1->25->24 [5] 26/-1/-1->25->24 [6] 26/-1/-1->25->24 [7] 26/-1/-1->25->24 [8] 26/-1/-1->25->24 [9] 26/9/-1->25->-1 [10] -1/-1/-1->25->24 [11] 26/-1/-1->25->24 [12] 26/-1/-1->25->24 [13] 26/-1/-1->25->24 [14] 26/-1/-1->25->24 [15] 26/-1/-1->25->24 +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Trees [0] 29/-1/-1->28->27 [1] 29/-1/-1->28->27 [2] 29/-1/-1->28->27 [3] 29/-1/-1->28->27 [4] 29/-1/-1->28->20 [5] -1/-1/-1->28->27 [6] 29/-1/-1->28->27 [7] 29/-1/-1->28->27 [8] 29/-1/-1->28->27 [9] 29/-1/-1->28->27 [10] 29/-1/-1->28->27 [11] 29/-1/-1->28->27 [12] 29/12/-1->28->-1 [13] -1/-1/-1->28->27 [14] 29/-1/-1->28->27 [15] 29/-1/-1->28->27 +ip-26-0-163-58:940529:940858 [0] NCCL INFO Trees [0] 17/8/24->16->0 [1] -1/-1/-1->16->23 [2] 17/-1/-1->16->23 [3] 17/-1/-1->16->23 [4] 17/-1/-1->16->23 [5] 17/-1/-1->16->23 [6] 17/-1/-1->16->23 [7] 17/-1/-1->16->23 [8] 17/-1/-1->16->8 [9] -1/-1/-1->16->23 [10] 17/-1/-1->16->23 [11] 17/-1/-1->16->23 [12] 17/-1/-1->16->23 [13] 17/-1/-1->16->23 [14] 17/-1/-1->16->23 [15] 17/-1/-1->16->23 +ip-26-0-163-58:940529:940858 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940534:940860 [5] NCCL INFO Trees [0] 22/-1/-1->21->20 [1] 22/-1/-1->21->20 [2] 22/-1/-1->21->20 [3] 22/-1/-1->21->20 [4] 22/-1/-1->21->20 [5] 22/13/29->21->5 [6] -1/-1/-1->21->20 [7] 22/-1/-1->21->20 [8] 22/-1/-1->21->20 [9] 22/-1/-1->21->20 [10] 22/-1/-1->21->20 [11] 22/-1/-1->21->20 [12] 22/-1/-1->21->20 [13] 22/-1/-1->21->13 [14] -1/-1/-1->21->20 [15] 22/-1/-1->21->20 +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->18 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/18/2->10->26 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Trees [0] 28/-1/-1->27->26 [1] 28/-1/-1->27->26 [2] 28/-1/-1->27->26 [3] 28/-1/-1->27->19 [4] -1/-1/-1->27->26 [5] 28/-1/-1->27->26 [6] 28/-1/-1->27->26 [7] 28/-1/-1->27->26 [8] 28/-1/-1->27->26 [9] 28/-1/-1->27->26 [10] 28/-1/-1->27->26 [11] 28/11/-1->27->-1 [12] -1/-1/-1->27->26 [13] 28/-1/-1->27->26 [14] 28/-1/-1->27->26 [15] 28/-1/-1->27->26 +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940534:940860 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940531:940862 [2] NCCL INFO Trees [0] 19/-1/-1->18->17 [1] 19/-1/-1->18->17 [2] 19/10/26->18->2 [3] -1/-1/-1->18->17 [4] 19/-1/-1->18->17 [5] 19/-1/-1->18->17 [6] 19/-1/-1->18->17 [7] 19/-1/-1->18->17 [8] 19/-1/-1->18->17 [9] 19/-1/-1->18->17 [10] 19/-1/-1->18->10 [11] -1/-1/-1->18->17 [12] 19/-1/-1->18->17 [13] 19/-1/-1->18->17 [14] 19/-1/-1->18->17 [15] 19/-1/-1->18->17 +ip-26-0-163-58:940531:940862 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Trees [0] 30/-1/-1->29->28 [1] 30/-1/-1->29->28 [2] 30/-1/-1->29->28 [3] 30/-1/-1->29->28 [4] 30/-1/-1->29->28 [5] 30/-1/-1->29->21 [6] -1/-1/-1->29->28 [7] 30/-1/-1->29->28 [8] 30/-1/-1->29->28 [9] 30/-1/-1->29->28 [10] 30/-1/-1->29->28 [11] 30/-1/-1->29->28 [12] 30/-1/-1->29->28 [13] 30/13/-1->29->-1 [14] -1/-1/-1->29->28 [15] 30/-1/-1->29->28 +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Trees [0] 9/-1/-1->8->16 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/16/0->8->24 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 01/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/19/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/18/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Trees [0] 25/-1/-1->24->16 [1] -1/-1/-1->24->31 [2] 25/-1/-1->24->31 [3] 25/-1/-1->24->31 [4] 25/-1/-1->24->31 [5] 25/-1/-1->24->31 [6] 25/-1/-1->24->31 [7] 25/-1/-1->24->31 [8] 25/8/-1->24->-1 [9] -1/-1/-1->24->31 [10] 25/-1/-1->24->31 [11] 25/-1/-1->24->31 [12] 25/-1/-1->24->31 [13] 25/-1/-1->24->31 [14] 25/-1/-1->24->31 [15] 25/-1/-1->24->31 +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Trees [0] 27/-1/-1->26->25 [1] 27/-1/-1->26->25 [2] 27/-1/-1->26->18 [3] -1/-1/-1->26->25 [4] 27/-1/-1->26->25 [5] 27/-1/-1->26->25 [6] 27/-1/-1->26->25 [7] 27/-1/-1->26->25 [8] 27/-1/-1->26->25 [9] 27/-1/-1->26->25 [10] 27/10/-1->26->-1 [11] -1/-1/-1->26->25 [12] 27/-1/-1->26->25 [13] 27/-1/-1->26->25 [14] 27/-1/-1->26->25 [15] 27/-1/-1->26->25 +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/17/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 03/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 04/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 05/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 06/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 07/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 09/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 11/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 12/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 13/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 14/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 15/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Trees [0] 1/16/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 08/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 08/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 08/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 08/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 08/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 08/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 08/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 08/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 05/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 08/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 08/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 06/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 05/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 13/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 06/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 05/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 14/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 06/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 13/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 14/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 08/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 08/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 05/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 13/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 14/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 06/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 13/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 14/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 08/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 08/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 05/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 13/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 05/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 13/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 05/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 13/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 06/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 06/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 14/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 14/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 07/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 15/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 06/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 06/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 14/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 14/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 07/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 15/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 07/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 07/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 01/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 15/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 15/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 02/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 01/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 03/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 04/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 02/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 03/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 09/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 04/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 10/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 09/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 11/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 10/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 12/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 11/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 12/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 01/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 01/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 04/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 05/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 06/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 04/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 04/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 05/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 10/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 05/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 07/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 06/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 06/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 09/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 07/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 07/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 10/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 09/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 10/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 11/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 10/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 11/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 12/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 11/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 12/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 12/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 13/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 06/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 13/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 14/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 13/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 14/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 15/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 14/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 14/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 15/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 15/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 04/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 05/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 06/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 06/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 05/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 07/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 06/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 05/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 10/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 06/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 05/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 14/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 11/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 05/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 14/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 13/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 14/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 12/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 13/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 13/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 13/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 14/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 09/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 15/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 13/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 07/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 07/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 15/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 07/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 15/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 02/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 03/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 15/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 04/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 05/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 09/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 10/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 11/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 12/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 13/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 01/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 02/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 04/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 02/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 09/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 10/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 03/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 04/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 12/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 03/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 05/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 04/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 04/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 01/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 05/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 12/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 09/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 01/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 07/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 06/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 10/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 02/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 09/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 07/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 11/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 04/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 06/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 12/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 09/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 09/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 10/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 07/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 13/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 10/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 11/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 11/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 09/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 12/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 12/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 10/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 12/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 13/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 15/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 11/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 14/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 14/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 03/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 15/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 15/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 04/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 05/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 06/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 01/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 04/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 07/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 07/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 09/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 11/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 09/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 10/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 11/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 11/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 06/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 12/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 12/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 07/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 13/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 15/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 09/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 14/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 10/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 15/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 11/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 14/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 15/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Connected all rings +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Connected all rings +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Connected all rings +ip-26-0-163-58:940530:940859 [1] NCCL INFO Connected all rings +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Connected all rings +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 04/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 05/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 06/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 07/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 10/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 11/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Connected all rings +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 12/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 04/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Connected all rings +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 05/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 13/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 14/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 06/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 15/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 07/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 10/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 04/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 11/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 05/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 06/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 12/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Connected all rings +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 07/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 13/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 09/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 14/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 11/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 15/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Connected all rings +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Connected all rings +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 12/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 13/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Connected all rings +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 14/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Connected all rings +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 04/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Connected all rings +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 05/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Connected all rings +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 06/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 15/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 05/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 07/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 09/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 11/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 06/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 07/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 12/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 13/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Connected all rings +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 13/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 05/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 14/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 15/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 14/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 04/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 06/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 15/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 05/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 07/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Connected all rings +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 09/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 06/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 05/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Connected all rings +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 10/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 07/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 06/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Connected all rings +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 11/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 09/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 07/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 13/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 10/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 13/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 14/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 04/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 14/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 12/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Connected all rings +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 05/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 05/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 15/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 04/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 13/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 15/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 06/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 06/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 05/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 14/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 04/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 07/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 07/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 07/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 15/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 06/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 04/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 09/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 09/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 09/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 06/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 10/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 10/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 10/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 07/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 12/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 11/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 11/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 09/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 13/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 12/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 13/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 10/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 14/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 13/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 11/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 14/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 15/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 15/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 12/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 04/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 15/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 14/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 05/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 07/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 09/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 10/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 15/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 06/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 09/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 10/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 11/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 12/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 04/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 13/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 14/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 11/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 05/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 12/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 14/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 07/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 09/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 15/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 10/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 11/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 12/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 13/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 15/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 04/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 05/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 06/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 00/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 00/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 08/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 09/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 08/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 10/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 11/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Connected all rings +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 12/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 13/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 08/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 09/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 14/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 08/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 09/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 00/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 03/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 05/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 05/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 00/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 00/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 08/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 06/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 01/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 04/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 01/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 03/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 13/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 07/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 02/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 08/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 02/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 05/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 05/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 08/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 03/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 12/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 00/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 03/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 06/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 08/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 11/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 04/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 04/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 04/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 13/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 07/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 05/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 13/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 08/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 14/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 06/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 05/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 08/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 15/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 07/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 12/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 06/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 02/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 11/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 09/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 02/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 07/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 08/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 13/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 09/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 10/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 14/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 10/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 15/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 10/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 08/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 11/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 11/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 10/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 12/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 12/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 13/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 13/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 14/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 14/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 15/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 15/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 06/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 08/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 06/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 14/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 08/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 14/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NVLS comm 0xa577ac0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Connected all trees +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO NVLS comm 0x9246d00 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Connected all trees +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO NVLS comm 0x9918980 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Connected all trees +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO NVLS comm 0x9ad7240 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Connected all trees +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NVLS comm 0x9e90ba0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Connected all trees +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO NVLS comm 0x87c9d20 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Connected all trees +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO NVLS comm 0xa5af2f0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Connected all trees +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NVLS comm 0x8b20d00 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Connected all trees +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO NVLS comm 0xa5f3ea0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Connected all trees +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NVLS comm 0xa378ac0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Connected all trees +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NVLS comm 0xa386880 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO NVLS comm 0x932f470 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Connected all trees +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NVLS comm 0xa27e270 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Connected all trees +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO NVLS comm 0x9881ce0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO NVLS comm 0x910a920 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NVLS comm 0xa43e800 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Connected all trees +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NVLS comm 0x9744c20 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940532:940861 [3] NCCL INFO Connected all trees +ip-26-0-163-58:940532:940861 [3] NCCL INFO NVLS comm 0x898fc20 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940534:940860 [5] NCCL INFO Connected all trees +ip-26-0-163-58:940534:940860 [5] NCCL INFO NVLS comm 0xa5fab60 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940533:940865 [4] NCCL INFO Connected all trees +ip-26-0-163-58:940533:940865 [4] NCCL INFO NVLS comm 0x8a5db40 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940531:940862 [2] NCCL INFO Connected all trees +ip-26-0-163-58:940531:940862 [2] NCCL INFO NVLS comm 0x9c3e4f0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940530:940859 [1] NCCL INFO Connected all trees +ip-26-0-163-58:940530:940859 [1] NCCL INFO NVLS comm 0x8f99900 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Connected all trees +ip-26-0-163-58:940536:940863 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO NVLS comm 0x88318a0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO NVLS comm 0x99d85c0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO NVLS comm 0x9796520 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940536:940863 [7] NCCL INFO NVLS comm 0x8b898b0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO NVLS comm 0x9a261a0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO NVLS comm 0xa51fec0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940529:940858 [0] NCCL INFO Connected all trees +ip-26-0-163-58:940529:940858 [0] NCCL INFO NVLS comm 0x8950e00 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO NVLS comm 0xa312fb0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO NVLS comm 0x9d70da0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940535:940864 [6] NCCL INFO Connected all trees +ip-26-0-163-58:940535:940864 [6] NCCL INFO NVLS comm 0xa3bdc70 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940533:940865 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940535:940864 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940531:940862 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940860 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940532:940861 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:940858 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940530:940859 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940536:940863 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967687:1968039 [1] NCCL INFO comm 0x9e90ba0 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-161-138:1967691:1968041 [5] NCCL INFO comm 0xa577ac0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-161-138:1967688:1968038 [2] NCCL INFO comm 0xa378ac0 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-161-138:1967686:1968037 [0] NCCL INFO comm 0x8b20d00 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-161-138:1967690:1968044 [4] NCCL INFO comm 0xa27e270 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-161-138:1967689:1968042 [3] NCCL INFO comm 0x9744c20 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-161-138:1967693:1968043 [7] NCCL INFO comm 0xa386880 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-161-138:1967692:1968040 [6] NCCL INFO comm 0xa43e800 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940533:940865 [4] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940533:940865 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940533:940865 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940533:940865 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940529:940858 [0] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940529:940858 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940529:940858 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940529:940858 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940531:940862 [2] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940531:940862 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940531:940862 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940531:940862 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940535:940864 [6] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940535:940864 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940535:940864 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940535:940864 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207363:2207692 [6] NCCL INFO comm 0x87c9d20 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-169-239:2207361:2207688 [4] NCCL INFO comm 0x9246d00 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-169-239:2207360:2207690 [3] NCCL INFO comm 0x9918980 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-169-239:2207364:2207687 [7] NCCL INFO comm 0x9881ce0 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-169-239:2207358:2207691 [1] NCCL INFO comm 0xa5af2f0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-169-239:2207357:2207693 [0] NCCL INFO comm 0xa5f3ea0 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-169-239:2207362:2207694 [5] NCCL INFO comm 0x932f470 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-169-239:2207359:2207689 [2] NCCL INFO comm 0x9ad7240 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-58:940536:940863 [7] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940536:940863 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940536:940863 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940536:940863 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940530:940859 [1] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940530:940859 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940530:940859 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940530:940859 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940534:940860 [5] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940534:940860 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940534:940860 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940534:940860 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940532:940861 [3] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940532:940861 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940532:940861 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-163-58:940532:940861 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685414:1685766 [0] NCCL INFO comm 0xa312fb0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-220:1685418:1685770 [4] NCCL INFO comm 0x99d85c0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-220:1685415:1685767 [1] NCCL INFO comm 0x9796520 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-220:1685421:1685771 [7] NCCL INFO comm 0xa51fec0 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-220:1685419:1685769 [5] NCCL INFO comm 0x910a920 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-220:1685417:1685768 [3] NCCL INFO comm 0x88318a0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-220:1685416:1685765 [2] NCCL INFO comm 0x9a261a0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-220:1685420:1685772 [6] NCCL INFO comm 0x9d70da0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-58:940531:940862 [2] NCCL INFO comm 0x9c3e4f0 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-58:940535:940864 [6] NCCL INFO comm 0xa3bdc70 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-58:940529:940858 [0] NCCL INFO comm 0x8950e00 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-58:940536:940863 [7] NCCL INFO comm 0x8b898b0 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-58:940533:940865 [4] NCCL INFO comm 0x8a5db40 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-58:940532:940861 [3] NCCL INFO comm 0x898fc20 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-58:940530:940859 [1] NCCL INFO comm 0x8f99900 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-163-58:940534:940860 [5] NCCL INFO comm 0xa5fab60 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xea96d08e83e9950d - Init COMPLETE +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO comm 0xa38d500 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfe080d9a89984b35 - Init START +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO comm 0x9ea5dc0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfe080d9a89984b35 - Init START +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO comm 0x8b339c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfe080d9a89984b35 - Init START +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO comm 0xa39d640 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfe080d9a89984b35 - Init START +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO comm 0xa452cd0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfe080d9a89984b35 - Init START +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO comm 0x9759e60 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfe080d9a89984b35 - Init START +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO comm 0xa58bcf0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfe080d9a89984b35 - Init START +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO comm 0xa295880 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfe080d9a89984b35 - Init START +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Using network Libfabric +ip-26-0-163-58:940535:940894 [6] NCCL INFO Using network Libfabric +ip-26-0-163-58:940534:940895 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:940536:940896 [7] NCCL INFO Using network Libfabric +ip-26-0-163-58:940532:940897 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:940533:940898 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:940530:940899 [1] NCCL INFO Using network Libfabric +ip-26-0-163-58:940531:940900 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:940536:940896 [7] NCCL INFO comm 0x8ba1140 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc32d65efd00ec428 - Init START +ip-26-0-163-58:940532:940897 [3] NCCL INFO comm 0x89a5910 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc32d65efd00ec428 - Init START +ip-26-0-163-58:940531:940900 [2] NCCL INFO comm 0x9c56ac0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc32d65efd00ec428 - Init START +ip-26-0-163-58:940535:940894 [6] NCCL INFO comm 0xa3d6b20 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc32d65efd00ec428 - Init START +ip-26-0-163-58:940530:940899 [1] NCCL INFO comm 0x8fb0000 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc32d65efd00ec428 - Init START +ip-26-0-163-58:940533:940898 [4] NCCL INFO comm 0x8a72920 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc32d65efd00ec428 - Init START +ip-26-0-163-58:940529:940893 [0] NCCL INFO comm 0x8968680 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc32d65efd00ec428 - Init START +ip-26-0-163-58:940534:940895 [5] NCCL INFO comm 0xa6131a0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc32d65efd00ec428 - Init START +ip-26-0-163-58:940531:940900 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940532:940897 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940529:940893 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940534:940895 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940536:940896 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940535:940894 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940530:940899 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940533:940898 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO comm 0x9930bb0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3e1afdaa51881d55 - Init START +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO comm 0xa60a7d0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3e1afdaa51881d55 - Init START +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO comm 0x9aebb40 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3e1afdaa51881d55 - Init START +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO comm 0xa5c5ce0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e1afdaa51881d55 - Init START +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO comm 0x98974a0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3e1afdaa51881d55 - Init START +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO comm 0x9346560 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3e1afdaa51881d55 - Init START +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO comm 0x925cb20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x3e1afdaa51881d55 - Init START +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO comm 0x87dea60 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x3e1afdaa51881d55 - Init START +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO comm 0xa536040 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x26826808715cfd46 - Init START +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO comm 0x88485b0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x26826808715cfd46 - Init START +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO comm 0xa32c120 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x26826808715cfd46 - Init START +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO comm 0x9a3b780 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x26826808715cfd46 - Init START +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO comm 0x97ab980 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x26826808715cfd46 - Init START +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO comm 0x9d87230 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x26826808715cfd46 - Init START +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO comm 0x99eefe0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x26826808715cfd46 - Init START +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO comm 0x9120040 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x26826808715cfd46 - Init START +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940533:940898 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940534:940895 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-163-58:940536:940896 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940536:940896 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-163-58:940530:940899 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:940530:940899 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-58:940529:940893 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-163-58:940535:940894 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940535:940894 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-163-58:940531:940900 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-58:940531:940900 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-163-58:940531:940900 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-163-58:940531:940900 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940535:940894 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940535:940894 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940534:940895 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940536:940896 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940533:940898 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-163-58:940534:940895 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940536:940896 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940532:940897 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-163-58:940533:940898 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940530:940899 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-163-58:940532:940897 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-163-58:940530:940899 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940529:940893 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Connected all rings +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Connected all rings +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Connected all rings +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Connected all rings +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Connected all rings +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Connected all rings +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Connected all rings +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Connected all rings +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Connected all rings +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Connected all rings +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Connected all rings +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Connected all rings +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Connected all rings +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Connected all rings +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Connected all rings +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Connected all rings +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Connected all rings +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Connected all rings +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940529:940893 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO NVLS comm 0xa32c120 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Connected all rings +ip-26-0-163-58:940536:940896 [7] NCCL INFO Connected all rings +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Connected all rings +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Connected all trees +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO NVLS comm 0xa39d640 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Connected all trees +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO NVLS comm 0x8b339c0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940535:940894 [6] NCCL INFO Connected all rings +ip-26-0-163-58:940533:940898 [4] NCCL INFO Connected all rings +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Connected all rings +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Connected all trees +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO NVLS comm 0x98974a0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940536:940896 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940895 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940532:940897 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940894 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940899 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Connected all trees +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO NVLS comm 0xa60a7d0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940531:940900 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-163-58:940533:940898 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Connected all trees +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO NVLS comm 0xa38d500 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Connected all trees +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO NVLS comm 0x9759e60 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Connected all trees +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO NVLS comm 0xa295880 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO NVLS comm 0xa536040 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Connected all trees +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO NVLS comm 0xa452cd0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Connected all trees +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO NVLS comm 0x9ea5dc0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO NVLS comm 0xa58bcf0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO NVLS comm 0x9a3b780 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Connected all trees +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO NVLS comm 0xa5c5ce0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940536:940896 [7] NCCL INFO Connected all trees +ip-26-0-163-58:940536:940896 [7] NCCL INFO NVLS comm 0x8ba1140 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Connected all trees +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO NVLS comm 0x9930bb0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO NVLS comm 0x97ab980 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Connected all trees +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO NVLS comm 0x9346560 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Connected all trees +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO NVLS comm 0x9aebb40 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-161-138:1967691:1968078 [5] NCCL INFO comm 0xa58bcf0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xfe080d9a89984b35 - Init COMPLETE +ip-26-0-161-138:1967687:1968080 [1] NCCL INFO comm 0x9ea5dc0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xfe080d9a89984b35 - Init COMPLETE +ip-26-0-161-138:1967689:1968081 [3] NCCL INFO comm 0x9759e60 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfe080d9a89984b35 - Init COMPLETE +ip-26-0-161-138:1967693:1968082 [7] NCCL INFO comm 0xa39d640 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xfe080d9a89984b35 - Init COMPLETE +ip-26-0-161-138:1967690:1968083 [4] NCCL INFO comm 0xa295880 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xfe080d9a89984b35 - Init COMPLETE +ip-26-0-161-138:1967686:1968077 [0] NCCL INFO comm 0x8b339c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xfe080d9a89984b35 - Init COMPLETE +ip-26-0-161-138:1967688:1968079 [2] NCCL INFO comm 0xa38d500 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfe080d9a89984b35 - Init COMPLETE +ip-26-0-161-138:1967692:1968084 [6] NCCL INFO comm 0xa452cd0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfe080d9a89984b35 - Init COMPLETE +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Connected all trees +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO NVLS comm 0x925cb20 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Connected all trees +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO NVLS comm 0x87dea60 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940529:940893 [0] NCCL INFO Connected all trees +ip-26-0-163-58:940529:940893 [0] NCCL INFO NVLS comm 0x8968680 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO NVLS comm 0x9d87230 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO NVLS comm 0x99eefe0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO NVLS comm 0x9120040 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO NVLS comm 0x88485b0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207364:2207725 [7] NCCL INFO comm 0x98974a0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3e1afdaa51881d55 - Init COMPLETE +ip-26-0-169-239:2207360:2207726 [3] NCCL INFO comm 0x9930bb0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x3e1afdaa51881d55 - Init COMPLETE +ip-26-0-169-239:2207358:2207727 [1] NCCL INFO comm 0xa5c5ce0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e1afdaa51881d55 - Init COMPLETE +ip-26-0-169-239:2207362:2207730 [5] NCCL INFO comm 0x9346560 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x3e1afdaa51881d55 - Init COMPLETE +ip-26-0-169-239:2207363:2207729 [6] NCCL INFO comm 0x87dea60 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x3e1afdaa51881d55 - Init COMPLETE +ip-26-0-169-239:2207361:2207728 [4] NCCL INFO comm 0x925cb20 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x3e1afdaa51881d55 - Init COMPLETE +ip-26-0-169-239:2207359:2207731 [2] NCCL INFO comm 0x9aebb40 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3e1afdaa51881d55 - Init COMPLETE +ip-26-0-169-239:2207357:2207724 [0] NCCL INFO comm 0xa60a7d0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x3e1afdaa51881d55 - Init COMPLETE +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685417:1685810 [3] NCCL INFO comm 0x88485b0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x26826808715cfd46 - Init COMPLETE +ip-26-0-163-220:1685421:1685807 [7] NCCL INFO comm 0xa536040 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x26826808715cfd46 - Init COMPLETE +ip-26-0-163-220:1685415:1685811 [1] NCCL INFO comm 0x97ab980 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x26826808715cfd46 - Init COMPLETE +ip-26-0-163-220:1685419:1685808 [5] NCCL INFO comm 0x9120040 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x26826808715cfd46 - Init COMPLETE +ip-26-0-163-220:1685416:1685809 [2] NCCL INFO comm 0x9a3b780 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x26826808715cfd46 - Init COMPLETE +ip-26-0-163-220:1685420:1685806 [6] NCCL INFO comm 0x9d87230 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x26826808715cfd46 - Init COMPLETE +ip-26-0-163-220:1685414:1685805 [0] NCCL INFO comm 0xa32c120 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x26826808715cfd46 - Init COMPLETE +ip-26-0-163-220:1685418:1685812 [4] NCCL INFO comm 0x99eefe0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x26826808715cfd46 - Init COMPLETE +ip-26-0-163-58:940535:940894 [6] NCCL INFO Connected all trees +ip-26-0-163-58:940535:940894 [6] NCCL INFO NVLS comm 0xa3d6b20 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940530:940899 [1] NCCL INFO Connected all trees +ip-26-0-163-58:940530:940899 [1] NCCL INFO NVLS comm 0x8fb0000 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940533:940898 [4] NCCL INFO Connected all trees +ip-26-0-163-58:940533:940898 [4] NCCL INFO NVLS comm 0x8a72920 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940534:940895 [5] NCCL INFO Connected all trees +ip-26-0-163-58:940534:940895 [5] NCCL INFO NVLS comm 0xa6131a0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940531:940900 [2] NCCL INFO Connected all trees +ip-26-0-163-58:940531:940900 [2] NCCL INFO NVLS comm 0x9c56ac0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940532:940897 [3] NCCL INFO Connected all trees +ip-26-0-163-58:940532:940897 [3] NCCL INFO NVLS comm 0x89a5910 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-163-58:940535:940894 [6] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940535:940894 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940535:940894 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940535:940894 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940529:940893 [0] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940529:940893 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940529:940893 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940529:940893 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940533:940898 [4] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940533:940898 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940533:940898 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940533:940898 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940531:940900 [2] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940531:940900 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940531:940900 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940531:940900 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940536:940896 [7] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940536:940896 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940536:940896 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940536:940896 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940532:940897 [3] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940530:940899 [1] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940532:940897 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940530:940899 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940532:940897 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940532:940897 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940530:940899 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940530:940899 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940534:940895 [5] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940534:940895 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940534:940895 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940534:940895 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940535:940894 [6] NCCL INFO comm 0xa3d6b20 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc32d65efd00ec428 - Init COMPLETE +ip-26-0-163-58:940531:940900 [2] NCCL INFO comm 0x9c56ac0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc32d65efd00ec428 - Init COMPLETE +ip-26-0-163-58:940529:940893 [0] NCCL INFO comm 0x8968680 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc32d65efd00ec428 - Init COMPLETE +ip-26-0-163-58:940533:940898 [4] NCCL INFO comm 0x8a72920 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc32d65efd00ec428 - Init COMPLETE +ip-26-0-163-58:940536:940896 [7] NCCL INFO comm 0x8ba1140 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc32d65efd00ec428 - Init COMPLETE +ip-26-0-163-58:940534:940895 [5] NCCL INFO comm 0xa6131a0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc32d65efd00ec428 - Init COMPLETE +ip-26-0-163-58:940532:940897 [3] NCCL INFO comm 0x89a5910 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc32d65efd00ec428 - Init COMPLETE +ip-26-0-163-58:940530:940899 [1] NCCL INFO comm 0x8fb0000 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc32d65efd00ec428 - Init COMPLETE +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Config: +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Config(general=GeneralArgs(project='debug', +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: run='1.34G_dp4_tp1_pp8_acc32_mbs2_seq4096_zero1_tpmodeRED_vocab131k', +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: seed=42, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: step=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: consumed_train_samples=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: ignore_sanity_checks=True), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: parallelism=ParallelismArgs(dp=4, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pp=8, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tp=1, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pp_engine=, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tp_mode=, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tp_linear_async_communication=True, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: recompute_layer=False, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tp_recompute_allgather=True, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: expert_parallel_size=1), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: eos_token_id=0, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: hidden_act='silu', +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: hidden_size=2048, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: initializer_range=0.02, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: intermediate_size=8192, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: is_llama_config=True, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: max_position_embeddings=4096, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_attention_heads=32, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_hidden_layers=16, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_key_value_heads=32, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pad_token_id=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pretraining_tp=1, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rms_norm_eps=1e-05, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_scaling=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_theta=10000.0, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_interleaved=False, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tie_word_embeddings=True, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: use_cache=True, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: vocab_size=131072), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: init_method=RandomInit(std=0.02), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: dtype=torch.bfloat16, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: make_vocab_size_divisible_by=1, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: ddp_bucket_cap_mb=25), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tokenizer_revision=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tokenizer_max_length=None), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: checkpoint_interval=10000, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: save_initial_state=False, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: save_final_state=False, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: resume_checkpoint_path=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: checkpoints_path_is_shared_file_system=False), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: logging=LoggingArgs(log_level='info', +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: log_level_replica='info', +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: iteration_step_info_interval=1), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tokens=TokensArgs(sequence_length=4096, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: train_steps=100, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: micro_batch_size=2, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: batch_accumulation_per_replica=32, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: val_check_interval=100, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: limit_val_batches=0, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: limit_test_batches=0), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: adam_beta1=0.9, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: adam_beta2=0.95, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: torch_adam_is_fused=True, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: name='adamW'), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: zero_stage=1, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: weight_decay=0.01, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: clip_grad=1.0, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: accumulate_grad_in_fp32=True, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lr_warmup_steps=2, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lr_warmup_style='linear', +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lr_decay_style='cosine', +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lr_decay_steps=13, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lr_decay_starting_step=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: min_decay_lr=1e-05)), +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: start_training_step=1, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: data=DataArgs(dataset=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: seed=42, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_loading_workers=1))], +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: profiler=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lighteval=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: s3_upload=None) +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Model Config: +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: LlamaConfig(bos_token_id=0, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: eos_token_id=0, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: hidden_act='silu', +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: hidden_size=2048, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: initializer_range=0.02, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: intermediate_size=8192, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: is_llama_config=True, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: max_position_embeddings=4096, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_attention_heads=32, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_hidden_layers=16, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_key_value_heads=32, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pad_token_id=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pretraining_tp=1, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rms_norm_eps=1e-05, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_scaling=None, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_theta=10000.0, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_interleaved=False, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tie_word_embeddings=True, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: use_cache=True, +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: vocab_size=131072) +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Building model.. +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Initialize RoPE Theta = 10000.0 +12/28/2024 01:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Using network Libfabric +ip-26-0-163-58:940530:940938 [1] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:940531:940940 [2] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:940534:940937 [5] NCCL INFO comm 0xb8e3890 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x58c5bf74ff91c65 - Init START +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO comm 0xb805740 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xafabaf2d7d393ea2 - Init START +ip-26-0-163-58:940530:940938 [1] NCCL INFO comm 0xa27f6d0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xcbe4a5a664d212e1 - Init START +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Using network Libfabric +ip-26-0-163-58:940529:940942 [0] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:940534:940937 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:940530:940938 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940535:940936 [6] NCCL INFO comm 0xb6a63b0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2982e4171bd34c0 - Init START +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO comm 0xa3ef830 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x730ad384c7b4334a - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:940531:940940 [2] NCCL INFO comm 0xaf25bf0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3dfad2d5554f9c7f - Init START +ip-26-0-163-58:940535:940936 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-163-58:940531:940940 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940529:940942 [0] NCCL INFO comm 0x9c38b30 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7de9d75346defd41 - Init START +ip-26-0-163-58:940536:940944 [7] NCCL INFO comm 0x9e70e20 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9e0e724ed7d1e875 - Init START +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO comm 0xb0569e0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x410ff4ba3bbc9e0e - Init START +ip-26-0-163-58:940532:940934 [3] NCCL INFO comm 0x9c74450 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x41fbc495da11cb74 - Init START +ip-26-0-163-58:940533:940943 [4] NCCL INFO comm 0x9d435b0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x84c4e2ad4dd70a1b - Init START +ip-26-0-163-58:940536:940944 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940529:940942 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940532:940934 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940533:940943 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO comm 0x9ed2d30 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa7b780c08a4b4f0e - Init START +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO comm 0x9b18680 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe41bcbad500bbf8a - Init START +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO comm 0xaa7baf0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xecc9fc0f4a77be21 - Init START +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO comm 0xb567710 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2ad3ed9b09b82b6e - Init START +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO comm 0xb725db0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd9efd7eb96237302 - Init START +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO comm 0xad0acc0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x17d39fbc2a0f594e - Init START +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO comm 0xacc1630 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x41f852ad722fe768 - Init START +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO comm 0x9899570 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x434d6db9322890e3 - Init START +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO comm 0xb66fb10 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3946dd18c9ab82d3 - Init START +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO comm 0xb85cf90 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x542cce06d30cdbb5 - Init START +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO comm 0xb5fbd00 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7a1aa29594560f20 - Init START +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO comm 0x8e1b770 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe2dae3bf39c4c188 - Init START +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO comm 0xa3c8040 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeae17dad1e5b6833 - Init START +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO comm 0xb823450 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x20463df959380d2d - Init START +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO comm 0xab49de0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1d99606782781638 - Init START +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO comm 0xb3391e0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x27ebe16bbede5faa - Init START +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO comm 0x9fc8e60 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1192631ff5e95be3 - Init START +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO comm 0xad05550 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x23275a1e348a6d13 - Init START +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO comm 0xc267000 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc2102f5f71fcc3fb - Init START +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO comm 0xb7dd880 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9300ec9c2a15e7c7 - Init START +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO comm 0xabee9c0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb8f031f971978f45 - Init START +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940529:940942 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-58:940529:940942 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940529:940942 [0] NCCL INFO Connected all rings +ip-26-0-163-58:940529:940942 [0] NCCL INFO Connected all trees +ip-26-0-163-58:940529:940942 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Connected all rings +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO Connected all trees +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685420:1685846 [6] NCCL INFO comm 0xb0569e0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x410ff4ba3bbc9e0e - Init COMPLETE +ip-26-0-163-58:940536:940944 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-58:940536:940944 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940536:940944 [7] NCCL INFO Connected all rings +ip-26-0-163-58:940536:940944 [7] NCCL INFO Connected all trees +ip-26-0-163-58:940536:940944 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940532:940934 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-58:940532:940934 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940532:940934 [3] NCCL INFO Connected all rings +ip-26-0-163-58:940532:940934 [3] NCCL INFO Connected all trees +ip-26-0-163-58:940532:940934 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940529:940942 [0] NCCL INFO comm 0x9c38b30 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7de9d75346defd41 - Init COMPLETE +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Connected all rings +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Connected all rings +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO Connected all trees +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207361:2207767 [4] NCCL INFO comm 0x9899570 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x434d6db9322890e3 - Init COMPLETE +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940531:940940 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-58:940531:940940 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940531:940940 [2] NCCL INFO Connected all rings +ip-26-0-163-58:940531:940940 [2] NCCL INFO Connected all trees +ip-26-0-163-58:940531:940940 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Connected all rings +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO Connected all trees +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940536:940944 [7] NCCL INFO comm 0x9e70e20 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x9e0e724ed7d1e875 - Init COMPLETE +ip-26-0-163-58:940532:940934 [3] NCCL INFO comm 0x9c74450 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x41fbc495da11cb74 - Init COMPLETE +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Connected all rings +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940530:940938 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-58:940530:940938 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940530:940938 [1] NCCL INFO Connected all rings +ip-26-0-163-58:940530:940938 [1] NCCL INFO Connected all trees +ip-26-0-163-58:940530:940938 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Connected all rings +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO Connected all trees +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685418:1685855 [4] NCCL INFO comm 0xacc1630 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x41f852ad722fe768 - Init COMPLETE +ip-26-0-163-58:940534:940937 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Connected all rings +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-58:940534:940937 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-58:940534:940937 [5] NCCL INFO Connected all rings +ip-26-0-163-58:940534:940937 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940534:940937 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Connected all rings +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Connected all rings +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO Connected all trees +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Connected all rings +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO Connected all trees +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685417:1685852 [3] NCCL INFO comm 0x9b18680 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xe41bcbad500bbf8a - Init COMPLETE +ip-26-0-169-239:2207364:2207763 [7] NCCL INFO comm 0x9ed2d30 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa7b780c08a4b4f0e - Init COMPLETE +ip-26-0-163-58:940535:940936 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-58:940535:940936 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940535:940936 [6] NCCL INFO Connected all rings +ip-26-0-163-58:940535:940936 [6] NCCL INFO Connected all trees +ip-26-0-163-58:940535:940936 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940531:940940 [2] NCCL INFO comm 0xaf25bf0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3dfad2d5554f9c7f - Init COMPLETE +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685414:1685857 [0] NCCL INFO comm 0xb5fbd00 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7a1aa29594560f20 - Init COMPLETE +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Connected all rings +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO Connected all trees +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Connected all rings +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO Connected all trees +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685421:1685845 [7] NCCL INFO comm 0xb805740 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xafabaf2d7d393ea2 - Init COMPLETE +ip-26-0-163-220:1685416:1685853 [2] NCCL INFO comm 0xad0acc0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x17d39fbc2a0f594e - Init COMPLETE +ip-26-0-163-58:940533:940943 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-163-220:1685419:1685847 [5] NCCL INFO comm 0xa3ef830 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x730ad384c7b4334a - Init COMPLETE +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-163-58:940533:940943 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-163-58:940533:940943 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685415:1685850 [1] NCCL INFO comm 0xaa7baf0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xecc9fc0f4a77be21 - Init COMPLETE +ip-26-0-163-58:940530:940938 [1] NCCL INFO comm 0xa27f6d0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xcbe4a5a664d212e1 - Init COMPLETE +ip-26-0-163-58:940533:940943 [4] NCCL INFO Connected all rings +ip-26-0-163-58:940533:940943 [4] NCCL INFO Connected all trees +ip-26-0-163-58:940533:940943 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940534:940937 [5] NCCL INFO comm 0xb8e3890 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x58c5bf74ff91c65 - Init COMPLETE +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO Connected all trees +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207362:2207768 [5] NCCL INFO comm 0xa3c8040 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xeae17dad1e5b6833 - Init COMPLETE +ip-26-0-163-58:940535:940936 [6] NCCL INFO comm 0xb6a63b0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2982e4171bd34c0 - Init COMPLETE +ip-26-0-161-138:1967688:1968144 [2] NCCL INFO comm 0xc267000 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc2102f5f71fcc3fb - Init COMPLETE +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:940533:940943 [4] NCCL INFO comm 0x9d435b0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x84c4e2ad4dd70a1b - Init COMPLETE +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207363:2207769 [6] NCCL INFO comm 0x8e1b770 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe2dae3bf39c4c188 - Init COMPLETE +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207360:2207787 [3] NCCL INFO comm 0xab49de0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1d99606782781638 - Init COMPLETE +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO Connected all trees +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Connected all rings +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO Connected all trees +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Connected all rings +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO Connected all trees +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207359:2207786 [2] NCCL INFO comm 0xad05550 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x23275a1e348a6d13 - Init COMPLETE +ip-26-0-169-239:2207357:2207785 [0] NCCL INFO comm 0xb823450 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x20463df959380d2d - Init COMPLETE +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Connected all rings +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967689:1968146 [3] NCCL INFO comm 0xabee9c0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb8f031f971978f45 - Init COMPLETE +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-169-239:2207358:2207789 [1] NCCL INFO comm 0xb7dd880 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x9300ec9c2a15e7c7 - Init COMPLETE +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO Connected all trees +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Connected all rings +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO Connected all trees +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Connected all rings +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO Connected all trees +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967686:1968142 [0] NCCL INFO comm 0x9fc8e60 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1192631ff5e95be3 - Init COMPLETE +ip-26-0-161-138:1967693:1968126 [7] NCCL INFO comm 0xb66fb10 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0x3946dd18c9ab82d3 - Init COMPLETE +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:940532:940963 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:940536:940962 [7] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967691:1968124 [5] NCCL INFO comm 0xb85cf90 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x542cce06d30cdbb5 - Init COMPLETE +ip-26-0-161-138:1967687:1968141 [1] NCCL INFO comm 0xb3391e0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x27ebe16bbede5faa - Init COMPLETE +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO comm 0xa917c30 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98948601662fff03 - Init START +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO comm 0xb634df0 rank 0 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98948601662fff03 - Init START +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO comm 0xc0b5960 rank 1 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98948601662fff03 - Init START +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO comm 0xc24b3d0 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98948601662fff03 - Init START +ip-26-0-163-58:940536:940962 [7] NCCL INFO comm 0xa8b6670 rank 5 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98948601662fff03 - Init START +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO comm 0xab54f30 rank 6 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98948601662fff03 - Init START +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO comm 0xa55e080 rank 2 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98948601662fff03 - Init START +ip-26-0-163-58:940532:940963 [3] NCCL INFO comm 0xa6b9ea0 rank 4 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98948601662fff03 - Init START +ip-26-0-163-58:940536:940962 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940532:940963 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:940534:940964 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Using network Libfabric +ip-26-0-163-58:940530:940965 [1] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO comm 0xc2a2ca0 rank 1 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc147d4532202f243 - Init START +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO comm 0xa3d2f50 rank 7 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc147d4532202f243 - Init START +ip-26-0-163-58:940534:940964 [5] NCCL INFO comm 0xc328da0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc147d4532202f243 - Init START +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO comm 0xae35520 rank 3 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc147d4532202f243 - Init START +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO comm 0xb4c1410 rank 2 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc147d4532202f243 - Init START +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO comm 0xbd7f760 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc147d4532202f243 - Init START +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO comm 0xb7e7ff0 rank 6 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc147d4532202f243 - Init START +ip-26-0-163-58:940530:940965 [1] NCCL INFO comm 0xacc50e0 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc147d4532202f243 - Init START +ip-26-0-161-138:1967692:1968123 [6] NCCL INFO comm 0xb725db0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd9efd7eb96237302 - Init COMPLETE +ip-26-0-163-58:940534:940964 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940530:940965 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967690:1968125 [4] NCCL INFO comm 0xb567710 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2ad3ed9b09b82b6e - Init COMPLETE +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Using network Libfabric +ip-26-0-163-58:940531:940966 [2] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:940535:940967 [6] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO comm 0xc16b960 rank 1 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4fd0a6a0ce78785f - Init START +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO comm 0xc271d90 rank 0 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4fd0a6a0ce78785f - Init START +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO comm 0x985fb60 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4fd0a6a0ce78785f - Init START +ip-26-0-163-58:940535:940967 [6] NCCL INFO comm 0xc0ebfb0 rank 5 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4fd0a6a0ce78785f - Init START +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO comm 0xba9c680 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4fd0a6a0ce78785f - Init START +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO comm 0xb7503b0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4fd0a6a0ce78785f - Init START +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO comm 0xad0ff80 rank 6 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4fd0a6a0ce78785f - Init START +ip-26-0-163-58:940531:940966 [2] NCCL INFO comm 0xb96b330 rank 4 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4fd0a6a0ce78785f - Init START +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Using network Libfabric +ip-26-0-163-58:940533:940969 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:940529:940968 [0] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940535:940967 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940531:940966 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO comm 0xaa0f0a0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94020cfd5e664c67 - Init START +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO comm 0xbfad2f0 rank 1 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x94020cfd5e664c67 - Init START +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO comm 0xa2ddea0 rank 7 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x94020cfd5e664c67 - Init START +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO comm 0xb707410 rank 3 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x94020cfd5e664c67 - Init START +ip-26-0-163-58:940533:940969 [4] NCCL INFO comm 0xa789100 rank 5 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x94020cfd5e664c67 - Init START +ip-26-0-163-58:940529:940968 [0] NCCL INFO comm 0xa67e1e0 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94020cfd5e664c67 - Init START +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO comm 0xb82ddc0 rank 6 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94020cfd5e664c67 - Init START +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO comm 0xc041600 rank 2 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94020cfd5e664c67 - Init START +ip-26-0-163-58:940533:940969 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940529:940968 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-58:940530:940965 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-163-58:940533:940969 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-58:940531:940966 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-58:940529:940968 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-58:940536:940962 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-58:940532:940963 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:940534:940964 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940535:940967 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 6/-1/-1->7->5 [3] 6/-1/-1->7->5 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] 6/3/-1->7->-1 [7] 6/3/-1->7->-1 +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 00/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 01/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 02/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 03/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 04/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/5/-1->1->-1 [3] 0/5/-1->1->-1 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->3 [7] 0/-1/-1->1->3 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 05/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 06/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 07/08 : 0 3 2 5 4 7 6 1 +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 2/-1/-1->3->5 [3] 2/-1/-1->3->5 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] 2/5/1->3->7 [7] 2/5/1->3->7 +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Trees [0] 3/-1/-1->2->4 [1] 3/-1/-1->2->4 [2] -1/-1/-1->2->3 [3] -1/-1/-1->2->3 [4] 3/4/0->2->6 [5] 3/4/0->2->6 [6] -1/-1/-1->2->3 [7] -1/-1/-1->2->3 +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Trees [0] 1/4/-1->0->-1 [1] 1/4/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 [4] 1/-1/-1->0->2 [5] 1/-1/-1->0->2 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940534:940964 [5] NCCL INFO Trees [0] -1/-1/-1->5->4 [1] -1/-1/-1->5->4 [2] 4/3/7->5->1 [3] 4/3/7->5->1 [4] -1/-1/-1->5->4 [5] -1/-1/-1->5->4 [6] 4/-1/-1->5->3 [7] 4/-1/-1->5->3 +ip-26-0-163-58:940534:940964 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Trees [0] 7/-1/-1->6->4 [1] 7/-1/-1->6->4 [2] -1/-1/-1->6->7 [3] -1/-1/-1->6->7 [4] 7/2/-1->6->-1 [5] 7/2/-1->6->-1 [6] -1/-1/-1->6->7 [7] -1/-1/-1->6->7 +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940530:940965 [1] NCCL INFO Trees [0] 5/2/6->4->0 [1] 5/2/6->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 [4] 5/-1/-1->4->2 [5] 5/-1/-1->4->2 [6] -1/-1/-1->4->5 [7] -1/-1/-1->4->5 +ip-26-0-163-58:940530:940965 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Trees [0] 7/-1/-1->6->4 [1] 7/-1/-1->6->4 [2] -1/-1/-1->6->7 [3] -1/-1/-1->6->7 [4] 7/2/-1->6->-1 [5] 7/2/-1->6->-1 [6] -1/-1/-1->6->7 [7] -1/-1/-1->6->7 +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 00/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 01/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 02/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/5/-1->1->-1 [3] 0/5/-1->1->-1 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->3 [7] 0/-1/-1->1->3 +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 6/-1/-1->7->5 [3] 6/-1/-1->7->5 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] 6/3/-1->7->-1 [7] 6/3/-1->7->-1 +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 2/-1/-1->3->5 [3] 2/-1/-1->3->5 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] 2/5/1->3->7 [7] 2/5/1->3->7 +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 03/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 04/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 05/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 06/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 07/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Trees [0] 1/4/-1->0->-1 [1] 1/4/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 [4] 1/-1/-1->0->2 [5] 1/-1/-1->0->2 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940536:940962 [7] NCCL INFO Trees [0] -1/-1/-1->5->4 [1] -1/-1/-1->5->4 [2] 4/3/7->5->1 [3] 4/3/7->5->1 [4] -1/-1/-1->5->4 [5] -1/-1/-1->5->4 [6] 4/-1/-1->5->3 [7] 4/-1/-1->5->3 +ip-26-0-163-58:940536:940962 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940532:940963 [3] NCCL INFO Trees [0] 5/2/6->4->0 [1] 5/2/6->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 [4] 5/-1/-1->4->2 [5] 5/-1/-1->4->2 [6] -1/-1/-1->4->5 [7] -1/-1/-1->4->5 +ip-26-0-163-58:940532:940963 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Trees [0] 3/-1/-1->2->4 [1] 3/-1/-1->2->4 [2] -1/-1/-1->2->3 [3] -1/-1/-1->2->3 [4] 3/4/0->2->6 [5] 3/4/0->2->6 [6] -1/-1/-1->2->3 [7] -1/-1/-1->2->3 +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Trees [0] 7/-1/-1->6->4 [1] 7/-1/-1->6->4 [2] -1/-1/-1->6->7 [3] -1/-1/-1->6->7 [4] 7/2/-1->6->-1 [5] 7/2/-1->6->-1 [6] -1/-1/-1->6->7 [7] -1/-1/-1->6->7 +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 00/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 01/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 02/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 03/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 04/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 05/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/5/-1->1->-1 [3] 0/5/-1->1->-1 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->3 [7] 0/-1/-1->1->3 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 06/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 07/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Trees [0] 1/4/-1->0->-1 [1] 1/4/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 [4] 1/-1/-1->0->2 [5] 1/-1/-1->0->2 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 2/-1/-1->3->5 [3] 2/-1/-1->3->5 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] 2/5/1->3->7 [7] 2/5/1->3->7 +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 6/-1/-1->7->5 [3] 6/-1/-1->7->5 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] 6/3/-1->7->-1 [7] 6/3/-1->7->-1 +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940529:940968 [0] NCCL INFO Trees [0] 5/2/6->4->0 [1] 5/2/6->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 [4] 5/-1/-1->4->2 [5] 5/-1/-1->4->2 [6] -1/-1/-1->4->5 [7] -1/-1/-1->4->5 +ip-26-0-163-58:940533:940969 [4] NCCL INFO Trees [0] -1/-1/-1->5->4 [1] -1/-1/-1->5->4 [2] 4/3/7->5->1 [3] 4/3/7->5->1 [4] -1/-1/-1->5->4 [5] -1/-1/-1->5->4 [6] 4/-1/-1->5->3 [7] 4/-1/-1->5->3 +ip-26-0-163-58:940529:940968 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940533:940969 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Trees [0] 3/-1/-1->2->4 [1] 3/-1/-1->2->4 [2] -1/-1/-1->2->3 [3] -1/-1/-1->2->3 [4] 3/4/0->2->6 [5] 3/4/0->2->6 [6] -1/-1/-1->2->3 [7] -1/-1/-1->2->3 +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] 6/-1/-1->7->5 [3] 6/-1/-1->7->5 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] 6/3/-1->7->-1 [7] 6/3/-1->7->-1 +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Trees [0] 7/-1/-1->6->4 [1] 7/-1/-1->6->4 [2] -1/-1/-1->6->7 [3] -1/-1/-1->6->7 [4] 7/2/-1->6->-1 [5] 7/2/-1->6->-1 [6] -1/-1/-1->6->7 [7] -1/-1/-1->6->7 +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/5/-1->1->-1 [3] 0/5/-1->1->-1 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->3 [7] 0/-1/-1->1->3 +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 00/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 01/08 : 0 3 2 5 4 7 6 1 +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Trees [0] 3/-1/-1->2->4 [1] 3/-1/-1->2->4 [2] -1/-1/-1->2->3 [3] -1/-1/-1->2->3 [4] 3/4/0->2->6 [5] 3/4/0->2->6 [6] -1/-1/-1->2->3 [7] -1/-1/-1->2->3 +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 02/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 03/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 04/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 05/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 06/08 : 0 1 2 3 4 5 6 7 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 07/08 : 0 3 2 5 4 7 6 1 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Trees [0] 1/4/-1->0->-1 [1] 1/4/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 [4] 1/-1/-1->0->2 [5] 1/-1/-1->0->2 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 2/-1/-1->3->5 [3] 2/-1/-1->3->5 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] 2/5/1->3->7 [7] 2/5/1->3->7 +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940531:940966 [2] NCCL INFO Trees [0] 5/2/6->4->0 [1] 5/2/6->4->0 [2] -1/-1/-1->4->5 [3] -1/-1/-1->4->5 [4] 5/-1/-1->4->2 [5] 5/-1/-1->4->2 [6] -1/-1/-1->4->5 [7] -1/-1/-1->4->5 +ip-26-0-163-58:940531:940966 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940535:940967 [6] NCCL INFO Trees [0] -1/-1/-1->5->4 [1] -1/-1/-1->5->4 [2] 4/3/7->5->1 [3] 4/3/7->5->1 [4] -1/-1/-1->5->4 [5] -1/-1/-1->5->4 [6] 4/-1/-1->5->3 [7] 4/-1/-1->5->3 +ip-26-0-163-58:940535:940967 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 00/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 00/0 : 5[5] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 02/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 02/0 : 5[5] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 00/0 : 3[5] -> 4[1] [send] via NET/Libfabric/1(2)/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 04/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 00/0 : 7[5] -> 0[1] [send] via NET/Libfabric/1(6)/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 04/0 : 5[5] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 02/0 : 3[5] -> 4[1] [send] via NET/Libfabric/1(2)/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 06/0 : 5[5] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 06/0 : 1[5] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[5] via P2P/IPC +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 00/0 : 6[1] -> 7[5] via P2P/IPC +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 02/0 : 7[5] -> 0[1] [send] via NET/Libfabric/1(6)/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 04/0 : 3[5] -> 4[1] [send] via NET/Libfabric/1(2)/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[5] via P2P/IPC +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 02/0 : 6[1] -> 7[5] via P2P/IPC +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 06/0 : 3[5] -> 4[1] [send] via NET/Libfabric/1(2)/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 04/0 : 7[5] -> 0[1] [send] via NET/Libfabric/1(6)/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 00/0 : 5[7] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 02/0 : 5[7] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 04/0 : 2[1] -> 3[5] via P2P/IPC +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 04/0 : 6[1] -> 7[5] via P2P/IPC +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[3] [send] via NET/Libfabric/3(6)/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 06/0 : 7[5] -> 0[1] [send] via NET/Libfabric/1(6)/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 04/0 : 5[7] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[3] [send] via NET/Libfabric/3(6)/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 06/0 : 2[1] -> 3[5] via P2P/IPC +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 06/0 : 5[7] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 00/0 : 6[3] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 06/0 : 6[1] -> 7[5] via P2P/IPC +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[3] [send] via NET/Libfabric/3(6)/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 00/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 00/0 : 3[7] -> 4[3] [send] via NET/Libfabric/3(2)/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 02/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 02/0 : 3[7] -> 4[3] [send] via NET/Libfabric/3(2)/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 02/0 : 6[3] -> 7[7] via P2P/IPC +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 04/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[3] [send] via NET/Libfabric/3(6)/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 04/0 : 3[7] -> 4[3] [send] via NET/Libfabric/3(2)/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 06/0 : 1[7] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 06/0 : 3[7] -> 4[3] [send] via NET/Libfabric/3(2)/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 04/0 : 6[3] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 00/0 : 7[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[1] [send] via NET/Libfabric/1(0)/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 02/0 : 7[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[1] [send] via NET/Libfabric/1(0)/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 06/0 : 6[3] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 04/0 : 7[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 04/0 : 1[5] -> 2[1] [send] via NET/Libfabric/1(0)/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 06/0 : 7[5] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 06/0 : 1[5] -> 2[1] [send] via NET/Libfabric/1(0)/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 04/0 : 2[3] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 06/0 : 2[3] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 01/0 : 6[1] -> 1[5] [send] via NET/Libfabric/5(7)/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 03/0 : 6[1] -> 1[5] [send] via NET/Libfabric/5(7)/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 05/0 : 6[1] -> 1[5] [send] via NET/Libfabric/5(7)/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 07/0 : 6[1] -> 1[5] [send] via NET/Libfabric/5(7)/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 00/0 : 7[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[3] [send] via NET/Libfabric/3(0)/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 02/0 : 7[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[3] [send] via NET/Libfabric/3(0)/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 01/0 : 6[3] -> 1[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 04/0 : 7[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 03/0 : 6[3] -> 1[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 04/0 : 1[7] -> 2[3] [send] via NET/Libfabric/3(0)/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 06/0 : 7[7] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 05/0 : 6[3] -> 1[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 06/0 : 1[7] -> 2[3] [send] via NET/Libfabric/3(0)/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 07/0 : 6[3] -> 1[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 04/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 06/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 00/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 00/0 : 5[4] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 00/0 : 7[4] -> 0[0] [send] via NET/Libfabric/0(6)/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 02/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 02/0 : 5[4] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 00/0 : 3[4] -> 4[0] [send] via NET/Libfabric/0(2)/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 02/0 : 7[4] -> 0[0] [send] via NET/Libfabric/0(6)/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 04/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 04/0 : 5[4] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 02/0 : 3[4] -> 4[0] [send] via NET/Libfabric/0(2)/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 04/0 : 7[4] -> 0[0] [send] via NET/Libfabric/0(6)/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 06/0 : 1[4] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[4] via P2P/IPC +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 06/0 : 5[4] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 00/0 : 6[0] -> 7[4] via P2P/IPC +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 04/0 : 3[4] -> 4[0] [send] via NET/Libfabric/0(2)/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 06/0 : 7[4] -> 0[0] [send] via NET/Libfabric/0(6)/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 02/0 : 6[0] -> 7[4] via P2P/IPC +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[4] via P2P/IPC +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 06/0 : 3[4] -> 4[0] [send] via NET/Libfabric/0(2)/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 04/0 : 6[0] -> 7[4] via P2P/IPC +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 04/0 : 2[0] -> 3[4] via P2P/IPC +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 06/0 : 6[0] -> 7[4] via P2P/IPC +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 06/0 : 2[0] -> 3[4] via P2P/IPC +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 00/0 : 7[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 02/0 : 7[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 04/0 : 7[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 06/0 : 7[4] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 04/0 : 1[4] -> 2[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 06/0 : 1[4] -> 2[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 01/0 : 6[0] -> 1[4] [send] via NET/Libfabric/4(7)/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 03/0 : 6[0] -> 1[4] [send] via NET/Libfabric/4(7)/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 00/0 : 5[6] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 00/0 : 7[6] -> 0[2] [send] via NET/Libfabric/2(6)/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 02/0 : 5[6] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 02/0 : 7[6] -> 0[2] [send] via NET/Libfabric/2(6)/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 04/0 : 5[6] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 04/0 : 7[6] -> 0[2] [send] via NET/Libfabric/2(6)/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 06/0 : 5[6] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 00/0 : 6[2] -> 7[6] via P2P/IPC +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 06/0 : 7[6] -> 0[2] [send] via NET/Libfabric/2(6)/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 00/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 02/0 : 6[2] -> 7[6] via P2P/IPC +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 02/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 00/0 : 3[6] -> 4[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 04/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 04/0 : 6[2] -> 7[6] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 02/0 : 3[6] -> 4[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 06/0 : 1[6] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[6] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 04/0 : 3[6] -> 4[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 06/0 : 6[2] -> 7[6] via P2P/IPC +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[6] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 06/0 : 3[6] -> 4[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[6] via P2P/IPC +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[6] via P2P/IPC +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 01/0 : 6[2] -> 1[6] [send] via NET/Libfabric/6(7)/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 03/0 : 6[2] -> 1[6] [send] via NET/Libfabric/6(7)/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 05/0 : 6[2] -> 1[6] [send] via NET/Libfabric/6(7)/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 07/0 : 6[2] -> 1[6] [send] via NET/Libfabric/6(7)/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 00/0 : 3[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[1] [send] via NET/Libfabric/1(4)/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 02/0 : 3[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[1] [send] via NET/Libfabric/1(4)/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 04/0 : 3[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[1] [send] via NET/Libfabric/1(4)/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 06/0 : 3[5] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 00/0 : 4[1] -> 5[5] via P2P/IPC +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[1] [send] via NET/Libfabric/1(4)/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 02/0 : 4[1] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 04/0 : 4[1] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 06/0 : 4[1] -> 5[5] via P2P/IPC +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 00/0 : 3[7] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 00/0 : 5[7] -> 6[3] [send] via NET/Libfabric/3(4)/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 02/0 : 3[7] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 02/0 : 5[7] -> 6[3] [send] via NET/Libfabric/3(4)/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 04/0 : 3[7] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 04/0 : 5[7] -> 6[3] [send] via NET/Libfabric/3(4)/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 06/0 : 3[7] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 00/0 : 4[3] -> 5[7] via P2P/IPC +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 06/0 : 5[7] -> 6[3] [send] via NET/Libfabric/3(4)/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 02/0 : 4[3] -> 5[7] via P2P/IPC +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 04/0 : 4[3] -> 5[7] via P2P/IPC +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 06/0 : 4[3] -> 5[7] via P2P/IPC +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 00/0 : 3[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 02/0 : 3[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 00/0 : 5[4] -> 6[0] [send] via NET/Libfabric/0(4)/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 04/0 : 3[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 02/0 : 5[4] -> 6[0] [send] via NET/Libfabric/0(4)/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 06/0 : 3[4] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 00/0 : 4[0] -> 5[4] via P2P/IPC +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 04/0 : 5[4] -> 6[0] [send] via NET/Libfabric/0(4)/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 02/0 : 4[0] -> 5[4] via P2P/IPC +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 04/0 : 4[0] -> 5[4] via P2P/IPC +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 06/0 : 5[4] -> 6[0] [send] via NET/Libfabric/0(4)/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 06/0 : 4[0] -> 5[4] via P2P/IPC +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 01/0 : 4[0] -> 7[4] [send] via NET/Libfabric/4(5)/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 00/0 : 3[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 02/0 : 3[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 00/0 : 5[6] -> 6[2] [send] via NET/Libfabric/2(4)/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 04/0 : 3[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 02/0 : 5[6] -> 6[2] [send] via NET/Libfabric/2(4)/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 06/0 : 3[6] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 00/0 : 4[2] -> 5[6] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 04/0 : 5[6] -> 6[2] [send] via NET/Libfabric/2(4)/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 02/0 : 4[2] -> 5[6] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 06/0 : 5[6] -> 6[2] [send] via NET/Libfabric/2(4)/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 04/0 : 4[2] -> 5[6] via P2P/IPC +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 06/0 : 4[2] -> 5[6] via P2P/IPC +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 00/0 : 7[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[2] [send] via NET/Libfabric/2(0)/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 02/0 : 7[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[2] [send] via NET/Libfabric/2(0)/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 04/0 : 7[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 04/0 : 1[6] -> 2[2] [send] via NET/Libfabric/2(0)/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 06/0 : 7[6] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 06/0 : 1[6] -> 2[2] [send] via NET/Libfabric/2(0)/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 01/0 : 2[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 01/0 : 4[1] -> 7[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 03/0 : 2[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 01/0 : 4[1] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 03/0 : 4[1] -> 7[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 05/0 : 2[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 03/0 : 4[1] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 05/0 : 4[1] -> 7[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 05/0 : 4[1] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 07/0 : 2[1] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 07/0 : 4[1] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 07/0 : 4[1] -> 7[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 01/0 : 7[5] -> 6[1] via P2P/IPC +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 03/0 : 7[5] -> 6[1] via P2P/IPC +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 05/0 : 7[5] -> 6[1] via P2P/IPC +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 07/0 : 7[5] -> 6[1] via P2P/IPC +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 01/0 : 0[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 01/0 : 2[1] -> 5[5] [send] via NET/Libfabric/5(3)/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 03/0 : 0[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 03/0 : 2[1] -> 5[5] [send] via NET/Libfabric/5(3)/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 05/0 : 0[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 07/0 : 0[1] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 05/0 : 2[1] -> 5[5] [send] via NET/Libfabric/5(3)/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 07/0 : 2[1] -> 5[5] [send] via NET/Libfabric/5(3)/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[1] via P2P/IPC +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 01/0 : 2[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 01/0 : 4[3] -> 7[7] [send] via NET/Libfabric/7(5)/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 03/0 : 2[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 03/0 : 4[3] -> 7[7] [send] via NET/Libfabric/7(5)/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[1] via P2P/IPC +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 05/0 : 2[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 05/0 : 4[3] -> 7[7] [send] via NET/Libfabric/7(5)/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 07/0 : 2[3] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[1] via P2P/IPC +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 07/0 : 4[3] -> 7[7] [send] via NET/Libfabric/7(5)/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[1] via P2P/IPC +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 01/0 : 0[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 01/0 : 2[3] -> 5[7] [send] via NET/Libfabric/7(3)/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 03/0 : 0[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 03/0 : 2[3] -> 5[7] [send] via NET/Libfabric/7(3)/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 05/0 : 0[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 01/0 : 4[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 05/0 : 2[3] -> 5[7] [send] via NET/Libfabric/7(3)/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 03/0 : 4[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 07/0 : 0[3] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 07/0 : 2[3] -> 5[7] [send] via NET/Libfabric/7(3)/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 05/0 : 4[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 07/0 : 4[3] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 01/0 : 5[7] -> 4[3] via P2P/IPC +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[3] via P2P/IPC +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 01/0 : 6[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 01/0 : 0[1] -> 3[5] [send] via NET/Libfabric/5(1)/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 03/0 : 6[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[3] via P2P/IPC +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 03/0 : 0[1] -> 3[5] [send] via NET/Libfabric/5(1)/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 05/0 : 6[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 05/0 : 0[1] -> 3[5] [send] via NET/Libfabric/5(1)/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 07/0 : 6[1] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[3] via P2P/IPC +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 07/0 : 0[1] -> 3[5] [send] via NET/Libfabric/5(1)/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[3] via P2P/IPC +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[1] via P2P/IPC +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 03/0 : 3[5] -> 2[1] via P2P/IPC +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 05/0 : 3[5] -> 2[1] via P2P/IPC +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 01/0 : 6[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 01/0 : 0[3] -> 3[7] [send] via NET/Libfabric/7(1)/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 07/0 : 3[5] -> 2[1] via P2P/IPC +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 03/0 : 6[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 03/0 : 0[3] -> 3[7] [send] via NET/Libfabric/7(1)/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 05/0 : 6[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 05/0 : 0[3] -> 3[7] [send] via NET/Libfabric/7(1)/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 07/0 : 6[3] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 01/0 : 4[0] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 07/0 : 0[3] -> 3[7] [send] via NET/Libfabric/7(1)/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 05/0 : 6[0] -> 1[4] [send] via NET/Libfabric/4(7)/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 03/0 : 4[0] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 07/0 : 6[0] -> 1[4] [send] via NET/Libfabric/4(7)/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 05/0 : 4[0] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[3] via P2P/IPC +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 07/0 : 4[0] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 01/0 : 2[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 03/0 : 2[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 03/0 : 4[0] -> 7[4] [send] via NET/Libfabric/4(5)/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 05/0 : 2[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 05/0 : 4[0] -> 7[4] [send] via NET/Libfabric/4(5)/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 07/0 : 2[0] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 07/0 : 4[0] -> 7[4] [send] via NET/Libfabric/4(5)/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 01/0 : 7[4] -> 6[0] via P2P/IPC +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 03/0 : 5[7] -> 4[3] via P2P/IPC +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 03/0 : 7[4] -> 6[0] via P2P/IPC +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 01/0 : 0[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 01/0 : 2[0] -> 5[4] [send] via NET/Libfabric/4(3)/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 03/0 : 0[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 03/0 : 2[0] -> 5[4] [send] via NET/Libfabric/4(3)/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 05/0 : 0[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 05/0 : 2[0] -> 5[4] [send] via NET/Libfabric/4(3)/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 07/0 : 0[0] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 07/0 : 2[0] -> 5[4] [send] via NET/Libfabric/4(3)/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 01/0 : 5[4] -> 4[0] via P2P/IPC +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 05/0 : 7[4] -> 6[0] via P2P/IPC +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 05/0 : 5[7] -> 4[3] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 01/0 : 2[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 01/0 : 4[2] -> 7[6] [send] via NET/Libfabric/6(5)/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 03/0 : 2[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 03/0 : 4[2] -> 7[6] [send] via NET/Libfabric/6(5)/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 05/0 : 2[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 05/0 : 4[2] -> 7[6] [send] via NET/Libfabric/6(5)/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[3] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 07/0 : 2[2] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 07/0 : 4[2] -> 7[6] [send] via NET/Libfabric/6(5)/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 03/0 : 5[4] -> 4[0] via P2P/IPC +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 07/0 : 5[7] -> 4[3] via P2P/IPC +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 01/0 : 6[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 01/0 : 0[0] -> 3[4] [send] via NET/Libfabric/4(1)/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 03/0 : 6[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 07/0 : 7[4] -> 6[0] via P2P/IPC +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 03/0 : 0[0] -> 3[4] [send] via NET/Libfabric/4(1)/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 05/0 : 6[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 05/0 : 0[0] -> 3[4] [send] via NET/Libfabric/4(1)/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 07/0 : 6[0] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 07/0 : 0[0] -> 3[4] [send] via NET/Libfabric/4(1)/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 01/0 : 0[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[0] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 03/0 : 0[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 01/0 : 2[2] -> 5[6] [send] via NET/Libfabric/6(3)/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 05/0 : 0[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 03/0 : 2[2] -> 5[6] [send] via NET/Libfabric/6(3)/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 07/0 : 0[2] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 05/0 : 2[2] -> 5[6] [send] via NET/Libfabric/6(3)/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 07/0 : 2[2] -> 5[6] [send] via NET/Libfabric/6(3)/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 01/0 : 5[6] -> 4[2] via P2P/IPC +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 01/0 : 4[2] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 03/0 : 4[2] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 05/0 : 4[2] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 07/0 : 4[2] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 01/0 : 7[6] -> 6[2] via P2P/IPC +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[3] via P2P/IPC +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 03/0 : 3[4] -> 2[0] via P2P/IPC +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[3] via P2P/IPC +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 01/0 : 6[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 01/0 : 0[2] -> 3[6] [send] via NET/Libfabric/6(1)/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 03/0 : 6[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 03/0 : 0[2] -> 3[6] [send] via NET/Libfabric/6(1)/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 05/0 : 6[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 05/0 : 0[2] -> 3[6] [send] via NET/Libfabric/6(1)/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 07/0 : 6[2] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 07/0 : 0[2] -> 3[6] [send] via NET/Libfabric/6(1)/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 05/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 05/0 : 3[4] -> 2[0] via P2P/IPC +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 05/0 : 5[4] -> 4[0] via P2P/IPC +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 07/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 03/0 : 5[6] -> 4[2] via P2P/IPC +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 07/0 : 5[4] -> 4[0] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 05/0 : 5[6] -> 4[2] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 07/0 : 5[6] -> 4[2] via P2P/IPC +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 03/0 : 7[6] -> 6[2] via P2P/IPC +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 05/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 03/0 : 3[6] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 07/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 07/0 : 3[4] -> 2[0] via P2P/IPC +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 05/0 : 7[6] -> 6[2] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 05/0 : 3[6] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 07/0 : 7[6] -> 6[2] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 07/0 : 3[6] -> 2[2] via P2P/IPC +ip-26-0-163-58:940534:940964 [5] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Connected all rings +ip-26-0-163-58:940532:940963 [3] NCCL INFO Connected all rings +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 01/0 : 4[3] -> 5[7] via P2P/IPC +ip-26-0-163-58:940530:940965 [1] NCCL INFO Connected all rings +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 01/0 : 4[1] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 03/0 : 4[1] -> 5[5] via P2P/IPC +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[4] via P2P/IPC +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 05/0 : 4[1] -> 5[5] via P2P/IPC +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 07/0 : 4[1] -> 5[5] via P2P/IPC +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 02/0 : 3[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 00/0 : 2[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 03/0 : 3[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 01/0 : 2[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 04/0 : 2[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 06/0 : 3[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 05/0 : 2[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 07/0 : 3[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 00/0 : 4[1] -> 6[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 02/0 : 5[5] -> 7[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 01/0 : 4[1] -> 6[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 03/0 : 5[5] -> 7[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Connected all rings +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 03/0 : 4[3] -> 5[7] via P2P/IPC +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Connected all rings +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 05/0 : 4[3] -> 5[7] via P2P/IPC +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 07/0 : 4[3] -> 5[7] via P2P/IPC +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Connected all rings +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 05/0 : 2[3] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 07/0 : 2[3] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Connected all rings +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[5] via P2P/IPC +ip-26-0-163-58:940529:940968 [0] NCCL INFO Connected all rings +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 01/0 : 4[0] -> 5[4] via P2P/IPC +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 03/0 : 4[0] -> 5[4] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Connected all rings +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 05/0 : 4[0] -> 5[4] via P2P/IPC +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[5] via P2P/IPC +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 07/0 : 4[0] -> 5[4] via P2P/IPC +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 05/0 : 2[1] -> 3[5] via P2P/IPC +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 07/0 : 2[1] -> 3[5] via P2P/IPC +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 06/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 04/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 07/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 05/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 02/0 : 3[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 00/0 : 2[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 03/0 : 3[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 01/0 : 2[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 06/0 : 3[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 07/0 : 3[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 04/0 : 2[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 05/0 : 2[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 01/0 : 6[1] -> 7[5] via P2P/IPC +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 03/0 : 6[1] -> 7[5] via P2P/IPC +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 05/0 : 6[1] -> 7[5] via P2P/IPC +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 07/0 : 6[1] -> 7[5] via P2P/IPC +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Connected all rings +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 02/0 : 5[5] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 00/0 : 4[1] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 01/0 : 4[1] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 03/0 : 5[5] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 04/0 : 2[1] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 06/0 : 3[5] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 05/0 : 2[1] -> 6[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 07/0 : 3[5] -> 7[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 04/0 : 6[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 06/0 : 7[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 05/0 : 6[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 07/0 : 7[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 02/0 : 1[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 03/0 : 1[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 02/0 : 5[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 03/0 : 5[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[4] via P2P/IPC +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 00/0 : 2[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 02/0 : 3[4] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 01/0 : 2[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 03/0 : 3[4] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 04/0 : 2[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 06/0 : 3[4] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 05/0 : 2[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 07/0 : 3[4] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 00/0 : 4[0] -> 6[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 02/0 : 5[4] -> 7[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Connected all rings +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 01/0 : 4[0] -> 6[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 03/0 : 5[4] -> 7[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 05/0 : 2[0] -> 3[4] via P2P/IPC +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 01/0 : 6[0] -> 7[4] via P2P/IPC +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 07/0 : 2[0] -> 3[4] via P2P/IPC +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 03/0 : 6[0] -> 7[4] via P2P/IPC +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Connected all rings +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 05/0 : 6[0] -> 7[4] via P2P/IPC +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 05/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 07/0 : 6[0] -> 7[4] via P2P/IPC +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 07/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 04/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 06/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 02/0 : 5[4] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 05/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 07/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 00/0 : 4[0] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 03/0 : 5[4] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 01/0 : 4[0] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 04/0 : 2[0] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 06/0 : 3[4] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 05/0 : 2[0] -> 6[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 07/0 : 3[4] -> 7[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 04/0 : 6[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 06/0 : 7[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 05/0 : 6[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 07/0 : 7[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 04/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 06/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 07/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 05/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 00/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 02/0 : 5[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 04/0 : 6[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 06/0 : 7[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 01/0 : 4[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 03/0 : 5[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 00/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 02/0 : 1[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 07/0 : 7[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 05/0 : 6[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 06/0 : 3[5] -> 7[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 03/0 : 1[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 01/0 : 0[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 04/0 : 2[1] -> 6[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 07/0 : 3[5] -> 7[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 05/0 : 2[1] -> 6[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 06/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 04/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 07/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Channel 05/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 02/0 : 7[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 00/0 : 6[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 00/0 : 4[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 02/0 : 7[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 03/0 : 7[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 00/0 : 6[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 06/0 : 5[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 01/0 : 4[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 03/0 : 7[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 01/0 : 6[1] -> 4[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 07/0 : 5[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Channel 01/0 : 6[1] -> 4[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 00/0 : 4[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 02/0 : 5[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 04/0 : 4[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 06/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 03/0 : 5[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 01/0 : 4[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 05/0 : 4[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 07/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 04/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 06/0 : 5[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 04/0 : 4[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Channel 05/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 07/0 : 5[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940530:940965 [1] NCCL INFO Channel 05/0 : 4[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 00/0 : 7[5] -> 6[1] via P2P/IPC +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[1] via P2P/IPC +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[1] via P2P/IPC +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 04/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 06/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 07/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 05/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 02/0 : 3[4] -> 5[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 00/0 : 2[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 03/0 : 3[4] -> 5[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 01/0 : 2[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 06/0 : 3[4] -> 5[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 02/0 : 7[5] -> 6[1] via P2P/IPC +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 04/0 : 2[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 07/0 : 3[4] -> 5[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 05/0 : 2[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 02/0 : 1[4] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 04/0 : 7[5] -> 6[1] via P2P/IPC +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 03/0 : 1[4] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 02/0 : 5[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 03/0 : 5[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 02/0 : 3[5] -> 2[1] via P2P/IPC +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[1] via P2P/IPC +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 04/0 : 3[5] -> 2[1] via P2P/IPC +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 00/0 : 2[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 02/0 : 3[7] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[1] via P2P/IPC +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 01/0 : 2[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 03/0 : 3[7] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 06/0 : 3[7] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 04/0 : 2[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 06/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 07/0 : 3[7] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 05/0 : 2[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 07/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 02/0 : 5[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 00/0 : 4[3] -> 6[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 02/0 : 3[7] -> 5[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 03/0 : 3[7] -> 5[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 01/0 : 4[3] -> 6[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 03/0 : 5[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Channel 06/0 : 7[5] -> 6[1] via P2P/IPC +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Channel 06/0 : 3[5] -> 2[1] via P2P/IPC +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 06/0 : 3[7] -> 5[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 07/0 : 3[7] -> 5[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940534:940964 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[1] via P2P/IPC +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 02/0 : 5[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 03/0 : 5[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 06/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 02/0 : 1[7] -> 5[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 03/0 : 1[7] -> 5[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 06/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 07/0 : 3[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 04/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 05/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 00/0 : 2[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 01/0 : 2[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 04/0 : 2[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 05/0 : 2[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 00/0 : 4[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 01/0 : 4[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 04/0 : 6[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 00/0 : 0[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 05/0 : 6[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 01/0 : 0[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 04/0 : 2[3] -> 6[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 05/0 : 2[3] -> 6[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Connected all rings +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Connected all rings +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Connected all rings +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 01/0 : 6[3] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 06/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 04/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 03/0 : 6[3] -> 7[7] via P2P/IPC +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 07/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 05/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 05/0 : 6[3] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 07/0 : 6[3] -> 7[7] via P2P/IPC +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 02/0 : 5[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 00/0 : 4[3] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 01/0 : 4[3] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 03/0 : 5[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 04/0 : 2[3] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 06/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 00/0 : 0[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 05/0 : 2[3] -> 6[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 07/0 : 3[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 01/0 : 0[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 04/0 : 6[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 06/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 00/0 : 4[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 05/0 : 6[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 01/0 : 4[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 02/0 : 7[7] -> 5[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 00/0 : 6[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 03/0 : 7[7] -> 5[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 00/0 : 6[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Channel 01/0 : 6[3] -> 4[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 01/0 : 6[3] -> 4[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 00/0 : 4[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 01/0 : 4[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 02/0 : 1[7] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 04/0 : 4[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 03/0 : 1[7] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940532:940963 [3] NCCL INFO Channel 05/0 : 4[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 02/0 : 5[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 03/0 : 5[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Connected all rings +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 02/0 : 7[7] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 06/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 03/0 : 7[7] -> 5[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 07/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 02/0 : 5[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 03/0 : 5[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 06/0 : 5[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 07/0 : 5[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[3] via P2P/IPC +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 00/0 : 4[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 01/0 : 4[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 04/0 : 4[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 05/0 : 4[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 04/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Channel 05/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[3] via P2P/IPC +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 02/0 : 5[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 03/0 : 5[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 06/0 : 5[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 07/0 : 5[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 06/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 07/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[3] via P2P/IPC +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 00/0 : 5[7] -> 4[3] via P2P/IPC +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[3] via P2P/IPC +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[3] via P2P/IPC +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 04/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Channel 05/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[3] via P2P/IPC +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[3] via P2P/IPC +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Connected all rings +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 02/0 : 5[7] -> 4[3] via P2P/IPC +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[3] via P2P/IPC +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 04/0 : 5[7] -> 4[3] via P2P/IPC +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940536:940962 [7] NCCL INFO Channel 06/0 : 5[7] -> 4[3] via P2P/IPC +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 06/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 04/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 07/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 05/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 02/0 : 5[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 00/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 06/0 : 7[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 03/0 : 5[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 01/0 : 4[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 04/0 : 6[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 02/0 : 1[4] -> 5[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 07/0 : 7[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 00/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 03/0 : 1[4] -> 5[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 05/0 : 6[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 06/0 : 3[4] -> 7[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 01/0 : 0[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 04/0 : 2[0] -> 6[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 07/0 : 3[4] -> 7[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 05/0 : 2[0] -> 6[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 04/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 02/0 : 5[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Channel 05/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 00/0 : 4[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 03/0 : 5[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 02/0 : 7[4] -> 5[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 01/0 : 4[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 06/0 : 5[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 03/0 : 7[4] -> 5[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 00/0 : 6[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 04/0 : 4[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 07/0 : 5[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Channel 01/0 : 6[0] -> 4[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 05/0 : 4[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 06/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 07/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 04/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Channel 05/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 00/0 : 6[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 06/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 01/0 : 6[0] -> 4[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 07/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 00/0 : 4[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 01/0 : 4[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 04/0 : 4[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:940968 [0] NCCL INFO Channel 05/0 : 4[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 02/0 : 7[4] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 03/0 : 7[4] -> 5[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 02/0 : 5[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 03/0 : 5[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 06/0 : 5[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 07/0 : 5[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 00/0 : 7[4] -> 6[0] via P2P/IPC +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 00/0 : 5[4] -> 4[0] via P2P/IPC +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[0] via P2P/IPC +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Connected all rings +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 01/0 : 6[2] -> 7[6] via P2P/IPC +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 04/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 02/0 : 5[4] -> 4[0] via P2P/IPC +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Channel 06/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 02/0 : 3[4] -> 2[0] via P2P/IPC +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 04/0 : 5[4] -> 4[0] via P2P/IPC +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 02/0 : 7[4] -> 6[0] via P2P/IPC +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 04/0 : 3[4] -> 2[0] via P2P/IPC +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 04/0 : 7[4] -> 6[0] via P2P/IPC +ip-26-0-163-58:940533:940969 [4] NCCL INFO Channel 06/0 : 5[4] -> 4[0] via P2P/IPC +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Channel 06/0 : 7[4] -> 6[0] via P2P/IPC +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Channel 06/0 : 3[4] -> 2[0] via P2P/IPC +ip-26-0-163-58:940531:940966 [2] NCCL INFO Connected all rings +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 01/0 : 4[2] -> 5[6] via P2P/IPC +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 03/0 : 4[2] -> 5[6] via P2P/IPC +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 05/0 : 4[2] -> 5[6] via P2P/IPC +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 07/0 : 4[2] -> 5[6] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 02/0 : 3[6] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 03/0 : 3[6] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 06/0 : 3[6] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 04/0 : 2[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 07/0 : 3[6] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 05/0 : 2[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 02/0 : 5[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 00/0 : 4[2] -> 6[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 03/0 : 5[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 01/0 : 4[2] -> 6[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 03/0 : 6[2] -> 7[6] via P2P/IPC +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 05/0 : 6[2] -> 7[6] via P2P/IPC +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 07/0 : 6[2] -> 7[6] via P2P/IPC +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Connected all rings +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 00/0 : 4[2] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 02/0 : 5[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 03/0 : 5[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 01/0 : 4[2] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 06/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 04/0 : 2[2] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 07/0 : 3[6] -> 7[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 05/0 : 2[2] -> 6[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 06/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 04/0 : 6[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 07/0 : 7[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 05/0 : 6[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Connected all rings +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[6] via P2P/IPC +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[6] via P2P/IPC +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[6] via P2P/IPC +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[6] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 06/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 04/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 07/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 05/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 02/0 : 3[6] -> 5[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 00/0 : 2[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 03/0 : 3[6] -> 5[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 01/0 : 2[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 06/0 : 3[6] -> 5[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 04/0 : 2[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 07/0 : 3[6] -> 5[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 05/0 : 2[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 02/0 : 5[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 00/0 : 4[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 04/0 : 6[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 03/0 : 5[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 06/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 01/0 : 4[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 02/0 : 1[6] -> 5[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 00/0 : 0[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 05/0 : 6[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 07/0 : 7[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 04/0 : 2[2] -> 6[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 03/0 : 1[6] -> 5[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 00/0 : 0[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 02/0 : 1[6] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 06/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 01/0 : 0[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 01/0 : 0[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 05/0 : 2[2] -> 6[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 07/0 : 3[6] -> 7[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 03/0 : 1[6] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 00/0 : 4[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 02/0 : 5[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 01/0 : 4[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 00/0 : 4[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 03/0 : 5[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 02/0 : 5[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 00/0 : 6[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 01/0 : 4[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 03/0 : 5[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 02/0 : 7[6] -> 5[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 04/0 : 4[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 04/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 00/0 : 6[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 06/0 : 5[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 06/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 02/0 : 7[6] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 05/0 : 4[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Channel 05/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 01/0 : 6[2] -> 4[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 07/0 : 5[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 07/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 03/0 : 7[6] -> 5[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 04/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 00/0 : 4[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 06/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 02/0 : 5[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 01/0 : 4[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Channel 05/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 07/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 03/0 : 5[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 04/0 : 4[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 06/0 : 5[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940531:940966 [2] NCCL INFO Channel 05/0 : 4[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 07/0 : 5[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 03/0 : 7[6] -> 5[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 00/0 : 7[6] -> 6[2] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 00/0 : 5[6] -> 4[2] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 02/0 : 3[6] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 02/0 : 7[6] -> 6[2] via P2P/IPC +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 04/0 : 7[6] -> 6[2] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 02/0 : 5[6] -> 4[2] via P2P/IPC +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Channel 06/0 : 7[6] -> 6[2] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 04/0 : 3[6] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Channel 06/0 : 3[6] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 04/0 : 5[6] -> 4[2] via P2P/IPC +ip-26-0-163-58:940535:940967 [6] NCCL INFO Channel 06/0 : 5[6] -> 4[2] via P2P/IPC +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Channel 01/0 : 6[2] -> 4[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 04/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Channel 06/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO Connected all trees +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO Connected all trees +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967691:1968169 [5] NCCL INFO comm 0xc2a2ca0 rank 1 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc147d4532202f243 - Init COMPLETE +ip-26-0-161-138:1967687:1968168 [1] NCCL INFO comm 0xbd7f760 rank 0 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc147d4532202f243 - Init COMPLETE +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO Connected all trees +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO Connected all trees +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO Connected all trees +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO Connected all trees +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO Connected all trees +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO Connected all trees +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO Connected all trees +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO Connected all trees +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO Connected all trees +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207362:2207811 [5] NCCL INFO comm 0xa3d2f50 rank 7 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc147d4532202f243 - Init COMPLETE +ip-26-0-169-239:2207358:2207810 [1] NCCL INFO comm 0xb7e7ff0 rank 6 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc147d4532202f243 - Init COMPLETE +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO Connected all trees +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO Connected all trees +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO Connected all trees +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207364:2207808 [7] NCCL INFO comm 0xa917c30 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98948601662fff03 - Init COMPLETE +ip-26-0-169-239:2207360:2207809 [3] NCCL INFO comm 0xab54f30 rank 6 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98948601662fff03 - Init COMPLETE +ip-26-0-169-239:2207361:2207814 [4] NCCL INFO comm 0xa2ddea0 rank 7 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x94020cfd5e664c67 - Init COMPLETE +ip-26-0-169-239:2207357:2207815 [0] NCCL INFO comm 0xb82ddc0 rank 6 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94020cfd5e664c67 - Init COMPLETE +ip-26-0-161-138:1967692:1968172 [6] NCCL INFO comm 0xc16b960 rank 1 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4fd0a6a0ce78785f - Init COMPLETE +ip-26-0-161-138:1967688:1968171 [2] NCCL INFO comm 0xc271d90 rank 0 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4fd0a6a0ce78785f - Init COMPLETE +ip-26-0-161-138:1967693:1968165 [7] NCCL INFO comm 0xc0b5960 rank 1 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98948601662fff03 - Init COMPLETE +ip-26-0-161-138:1967689:1968164 [3] NCCL INFO comm 0xb634df0 rank 0 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98948601662fff03 - Init COMPLETE +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967690:1968175 [4] NCCL INFO comm 0xbfad2f0 rank 1 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x94020cfd5e664c67 - Init COMPLETE +ip-26-0-161-138:1967686:1968174 [0] NCCL INFO comm 0xaa0f0a0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94020cfd5e664c67 - Init COMPLETE +ip-26-0-163-58:940530:940965 [1] NCCL INFO Connected all trees +ip-26-0-163-58:940530:940965 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940530:940965 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940530:940965 [1] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940534:940964 [5] NCCL INFO Connected all trees +ip-26-0-163-58:940534:940964 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940534:940964 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940534:940964 [5] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940536:940962 [7] NCCL INFO Connected all trees +ip-26-0-163-58:940536:940962 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940536:940962 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940536:940962 [7] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940532:940963 [3] NCCL INFO Connected all trees +ip-26-0-163-58:940532:940963 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940532:940963 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940532:940963 [3] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685415:1685878 [1] NCCL INFO comm 0xb4c1410 rank 2 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc147d4532202f243 - Init COMPLETE +ip-26-0-163-220:1685419:1685877 [5] NCCL INFO comm 0xae35520 rank 3 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc147d4532202f243 - Init COMPLETE +ip-26-0-163-58:940529:940968 [0] NCCL INFO Connected all trees +ip-26-0-163-58:940529:940968 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940529:940968 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940529:940968 [0] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940533:940969 [4] NCCL INFO Connected all trees +ip-26-0-163-58:940533:940969 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940533:940969 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940533:940969 [4] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940530:940965 [1] NCCL INFO comm 0xacc50e0 rank 4 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc147d4532202f243 - Init COMPLETE +ip-26-0-163-58:940534:940964 [5] NCCL INFO comm 0xc328da0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc147d4532202f243 - Init COMPLETE +ip-26-0-163-220:1685417:1685876 [3] NCCL INFO comm 0xa55e080 rank 2 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98948601662fff03 - Init COMPLETE +ip-26-0-163-220:1685421:1685875 [7] NCCL INFO comm 0xc24b3d0 rank 3 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98948601662fff03 - Init COMPLETE +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO Connected all trees +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940536:940962 [7] NCCL INFO comm 0xa8b6670 rank 5 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x98948601662fff03 - Init COMPLETE +ip-26-0-163-58:940532:940963 [3] NCCL INFO comm 0xa6b9ea0 rank 4 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x98948601662fff03 - Init COMPLETE +ip-26-0-163-220:1685418:1685882 [4] NCCL INFO comm 0xb707410 rank 3 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x94020cfd5e664c67 - Init COMPLETE +ip-26-0-163-220:1685414:1685881 [0] NCCL INFO comm 0xc041600 rank 2 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94020cfd5e664c67 - Init COMPLETE +ip-26-0-169-239:2207363:2207813 [6] NCCL INFO comm 0x985fb60 rank 7 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4fd0a6a0ce78785f - Init COMPLETE +ip-26-0-169-239:2207359:2207812 [2] NCCL INFO comm 0xad0ff80 rank 6 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4fd0a6a0ce78785f - Init COMPLETE +ip-26-0-163-58:940529:940968 [0] NCCL INFO comm 0xa67e1e0 rank 4 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x94020cfd5e664c67 - Init COMPLETE +ip-26-0-163-58:940533:940969 [4] NCCL INFO comm 0xa789100 rank 5 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x94020cfd5e664c67 - Init COMPLETE +ip-26-0-163-58:940531:940966 [2] NCCL INFO Connected all trees +ip-26-0-163-58:940531:940966 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940531:940966 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940531:940966 [2] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940535:940967 [6] NCCL INFO Connected all trees +ip-26-0-163-58:940535:940967 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940535:940967 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-58:940535:940967 [6] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940531:940966 [2] NCCL INFO comm 0xb96b330 rank 4 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4fd0a6a0ce78785f - Init COMPLETE +ip-26-0-163-58:940535:940967 [6] NCCL INFO comm 0xc0ebfb0 rank 5 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4fd0a6a0ce78785f - Init COMPLETE +12/28/2024 01:47:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Total number of parameters: 1.61G (3072.13MiB) +12/28/2024 01:47:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Local number of parameters: 470M (896.02MiB) +12/28/2024 01:47:28 [INFO|DP=0|PP=1|TP=0|ip-26-0-161-138]: Local number of parameters: 201M (384.02MiB) +12/28/2024 01:47:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [After model building] Memory usage: 896.04MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +12/28/2024 01:47:28 [INFO|DP=0|PP=1|TP=0|ip-26-0-161-138]: [After model building] Memory usage: 384.04MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +12/28/2024 01:47:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: No checkpoint path provided. +12/28/2024 01:47:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Parametrizing model parameters using StandardParametrizator +12/28/2024 01:47:28 [INFO|DP=0|PP=7|TP=0|ip-26-0-169-239]: Local number of parameters: 0 (0.00MiB) +12/28/2024 01:47:28 [INFO|DP=0|PP=6|TP=0|ip-26-0-169-239]: Local number of parameters: 268M (512.00MiB) +12/28/2024 01:47:28 [INFO|DP=0|PP=5|TP=0|ip-26-0-163-58]: Local number of parameters: 201M (384.02MiB) +12/28/2024 01:47:28 [INFO|DP=0|PP=4|TP=0|ip-26-0-163-58]: Local number of parameters: 134M (256.02MiB) +12/28/2024 01:47:28 [INFO|DP=0|PP=7|TP=0|ip-26-0-169-239]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +12/28/2024 01:47:28 [INFO|DP=0|PP=5|TP=0|ip-26-0-163-58]: [After model building] Memory usage: 384.04MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +12/28/2024 01:47:28 [INFO|DP=0|PP=6|TP=0|ip-26-0-169-239]: [After model building] Memory usage: 512.01MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +12/28/2024 01:47:28 [INFO|DP=0|PP=4|TP=0|ip-26-0-163-58]: [After model building] Memory usage: 256.03MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +12/28/2024 01:47:28 [INFO|DP=0|PP=2|TP=0|ip-26-0-163-220]: Local number of parameters: 134M (256.02MiB) +12/28/2024 01:47:28 [INFO|DP=0|PP=3|TP=0|ip-26-0-163-220]: Local number of parameters: 201M (384.02MiB) +12/28/2024 01:47:28 [INFO|DP=0|PP=3|TP=0|ip-26-0-163-220]: [After model building] Memory usage: 384.04MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +12/28/2024 01:47:28 [INFO|DP=0|PP=2|TP=0|ip-26-0-163-220]: [After model building] Memory usage: 256.03MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO 8 coll channels, 0 nvls channels, 8 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685416:1685880 [2] NCCL INFO comm 0xb7503b0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4fd0a6a0ce78785f - Init COMPLETE +ip-26-0-163-220:1685420:1685879 [6] NCCL INFO comm 0xba9c680 rank 3 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4fd0a6a0ce78785f - Init COMPLETE +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO comm 0xb6516a0 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb94acd845dcd0d04 - Init START +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO comm 0xbd99ef0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb94acd845dcd0d04 - Init START +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO comm 0xaa2ef10 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb94acd845dcd0d04 - Init START +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO comm 0xc284a50 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb94acd845dcd0d04 - Init START +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO comm 0xc475ae0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5363b24ac14b9b1 - Init START +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO comm 0xc1816c0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5363b24ac14b9b1 - Init START +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO comm 0xc288e30 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5363b24ac14b9b1 - Init START +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO comm 0xc33dd00 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5363b24ac14b9b1 - Init START +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:940529:940992 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:940536:940996 [7] NCCL INFO Using network Libfabric +ip-26-0-163-58:940534:940997 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:940533:940993 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:940535:940995 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Using network Libfabric +ip-26-0-163-58:940532:940998 [3] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO comm 0xb9f6bf0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6032e321a0e84e9a - Init START +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO comm 0xad1ffd0 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6032e321a0e84e9a - Init START +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO comm 0xb9b3330 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6032e321a0e84e9a - Init START +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO comm 0xaeda130 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6032e321a0e84e9a - Init START +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940531:940994 [2] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO comm 0xc41f6a0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56732c84596033a6 - Init START +ip-26-0-163-58:940536:940996 [7] NCCL INFO comm 0xaa8c2a0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4251c4840ed1af9b - Init START +ip-26-0-163-58:940535:940995 [6] NCCL INFO comm 0xc2c1a00 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4251c4840ed1af9b - Init START +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO comm 0xbc6fc70 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56732c84596033a6 - Init START +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO comm 0xb8db680 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56732c84596033a6 - Init START +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO comm 0xb008df0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56732c84596033a6 - Init START +ip-26-0-163-58:940533:940993 [4] NCCL INFO comm 0xa95e6d0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4251c4840ed1af9b - Init START +ip-26-0-163-58:940534:940997 [5] NCCL INFO comm 0xc4fc660 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4251c4840ed1af9b - Init START +ip-26-0-163-58:940536:940996 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940535:940995 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940533:940993 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940534:940997 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940530:940999 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO comm 0xc214a80 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4c572e897855e7be - Init START +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO comm 0xa731c60 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4c572e897855e7be - Init START +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO comm 0xb694c70 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4c572e897855e7be - Init START +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO comm 0xb924060 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4c572e897855e7be - Init START +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940529:940992 [0] NCCL INFO comm 0xa8524c0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x126af73abfd8253c - Init START +ip-26-0-163-58:940530:940999 [1] NCCL INFO comm 0xae99da0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x126af73abfd8253c - Init START +ip-26-0-163-58:940532:940998 [3] NCCL INFO comm 0xa88e740 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x126af73abfd8253c - Init START +ip-26-0-163-58:940531:940994 [2] NCCL INFO comm 0xbb3ef50 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x126af73abfd8253c - Init START +ip-26-0-163-58:940532:940998 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940530:940999 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940531:940994 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940529:940992 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 04/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 05/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 06/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 07/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 08/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 09/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 10/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 11/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 12/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 13/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 14/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 15/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 16/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 17/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 18/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 19/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 20/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 21/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 22/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 23/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Connected all rings +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Connected all rings +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Connected all rings +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Connected all trees +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO NVLS comm 0xb9f6bf0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Connected all trees +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO NVLS comm 0xad1ffd0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Connected all trees +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO NVLS comm 0xaeda130 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Connected all trees +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO NVLS comm 0xb9b3330 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207360:2207839 [3] NCCL INFO comm 0xad1ffd0 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6032e321a0e84e9a - Init COMPLETE +ip-26-0-169-239:2207358:2207840 [1] NCCL INFO comm 0xb9b3330 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6032e321a0e84e9a - Init COMPLETE +ip-26-0-169-239:2207359:2207838 [2] NCCL INFO comm 0xaeda130 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x6032e321a0e84e9a - Init COMPLETE +ip-26-0-169-239:2207357:2207837 [0] NCCL INFO comm 0xb9f6bf0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x6032e321a0e84e9a - Init COMPLETE +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-58:940529:940992 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-163-58:940530:940999 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:940530:940999 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-163-58:940532:940998 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:940532:940998 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 04/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940533:940993 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-163-58:940536:940996 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940536:940996 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 04/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 05/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 04/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 06/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 05/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 07/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 05/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 06/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 08/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 06/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 07/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 04/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 09/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 07/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 05/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-58:940531:940994 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 06/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 08/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 10/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 08/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 07/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 08/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 09/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 11/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 09/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 09/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 10/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 10/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 12/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 10/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 11/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940534:940997 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940534:940997 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 12/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 11/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 11/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 13/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 12/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 12/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 13/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 14/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 15/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 13/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 16/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 13/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 14/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 17/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 14/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 18/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-163-58:940530:940999 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-163-58:940530:940999 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-163-58:940531:940994 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940532:940998 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-163-58:940532:940998 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 19/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 14/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-163-58:940529:940992 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-163-58:940529:940992 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 20/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 15/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940535:940995 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 15/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 16/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 16/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 21/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 17/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 17/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 15/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 22/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 18/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 18/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 16/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 23/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 17/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 18/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 19/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 20/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 21/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 19/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 19/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 22/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 20/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 20/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 23/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-163-58:940535:940995 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 21/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-163-58:940536:940996 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-163-58:940534:940997 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940536:940996 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-163-58:940533:940993 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-163-58:940533:940993 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 21/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 22/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 22/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 23/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 04/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 05/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 06/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 23/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 07/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 08/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 09/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 10/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 11/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 12/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 13/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 14/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 15/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 16/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 17/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 18/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 19/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 20/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 21/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 22/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 23/0 : 3[3] -> 0[0] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 04/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 04/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 05/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 05/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 06/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 07/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 06/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 08/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 07/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 09/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 04/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 08/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 10/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 05/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 11/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 09/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 06/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 12/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 10/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 07/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 13/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 08/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 11/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 14/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 12/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 04/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 09/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 15/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 13/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 16/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 14/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 17/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 15/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 05/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 10/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 18/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 16/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 19/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 17/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 06/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 11/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 20/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 12/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 18/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 21/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 19/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 22/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 20/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 23/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 21/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 07/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 13/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 22/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 08/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 14/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 23/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 09/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 15/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Connected all rings +ip-26-0-163-58:940531:940994 [2] NCCL INFO Connected all rings +ip-26-0-163-58:940529:940992 [0] NCCL INFO Connected all rings +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Connected all rings +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 10/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 16/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 11/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 17/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 18/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 12/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Connected all rings +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Connected all rings +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 19/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 13/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Connected all rings +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 14/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 20/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 15/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 21/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 16/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Connected all rings +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 22/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 17/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 23/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 18/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 19/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Connected all rings +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Connected all rings +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Connected all rings +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 20/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 21/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Connected all rings +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 22/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 23/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Connected all rings +ip-26-0-163-58:940536:940996 [7] NCCL INFO Connected all rings +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Connected all rings +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Connected all rings +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 04/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 05/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 06/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 07/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 08/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 09/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 10/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 11/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 12/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 13/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 14/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 15/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 16/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 17/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Connected all trees +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO NVLS comm 0xb8db680 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 04/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 18/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 19/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 20/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 05/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 21/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 22/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Channel 23/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 06/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO Connected all trees +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 07/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940529:940992 [0] NCCL INFO NVLS comm 0xa8524c0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 08/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940532:940998 [3] NCCL INFO Connected all trees +ip-26-0-163-58:940532:940998 [3] NCCL INFO NVLS comm 0xa88e740 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 09/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 10/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 11/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 12/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 13/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Connected all rings +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 14/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 15/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Connected all rings +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 16/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 17/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 18/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 19/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Connected all rings +ip-26-0-163-58:940531:940994 [2] NCCL INFO Connected all trees +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 20/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:940994 [2] NCCL INFO NVLS comm 0xbb3ef50 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 21/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Connected all rings +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:940997 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 22/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO Connected all trees +ip-26-0-163-58:940535:940995 [6] NCCL INFO Channel 23/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-163-58:940530:940999 [1] NCCL INFO NVLS comm 0xae99da0 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO NVLS comm 0xa731c60 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-163-58:940531:940994 [2] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940531:940994 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940531:940994 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:940531:940994 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940529:940992 [0] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940529:940992 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940529:940992 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:940529:940992 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940530:940999 [1] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940530:940999 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940530:940999 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:940530:940999 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940532:940998 [3] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940532:940998 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940532:940998 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:940532:940998 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO NVLS comm 0xc214a80 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Connected all trees +ip-26-0-163-58:940529:940992 [0] NCCL INFO comm 0xa8524c0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x126af73abfd8253c - Init COMPLETE +ip-26-0-163-58:940531:940994 [2] NCCL INFO comm 0xbb3ef50 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x126af73abfd8253c - Init COMPLETE +ip-26-0-163-58:940530:940999 [1] NCCL INFO comm 0xae99da0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x126af73abfd8253c - Init COMPLETE +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO NVLS comm 0xc41f6a0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-163-58:940532:940998 [3] NCCL INFO comm 0xa88e740 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x126af73abfd8253c - Init COMPLETE +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Connected all trees +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO NVLS comm 0xaa2ef10 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Connected all trees +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO NVLS comm 0xb6516a0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 04/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 05/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 06/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940536:940996 [7] NCCL INFO Connected all trees +ip-26-0-163-58:940536:940996 [7] NCCL INFO NVLS comm 0xaa8c2a0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 07/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 08/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 09/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 10/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-163-58:940533:940993 [4] NCCL INFO Connected all trees +ip-26-0-163-58:940533:940993 [4] NCCL INFO NVLS comm 0xa95e6d0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Connected all trees +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 11/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO NVLS comm 0xbd99ef0 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 12/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 13/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 14/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Connected all trees +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 15/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO NVLS comm 0xc284a50 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 16/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO NVLS comm 0xb694c70 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO NVLS comm 0xb924060 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-163-58:940535:940995 [6] NCCL INFO Connected all trees +ip-26-0-163-58:940535:940995 [6] NCCL INFO NVLS comm 0xc2c1a00 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-163-58:940534:940997 [5] NCCL INFO Connected all trees +ip-26-0-163-58:940534:940997 [5] NCCL INFO NVLS comm 0xc4fc660 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 17/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 18/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 19/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 20/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 21/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 22/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Channel 23/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO NVLS comm 0xb008df0 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO NVLS comm 0xbc6fc70 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940535:940995 [6] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940535:940995 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940535:940995 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:940535:940995 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940534:940997 [5] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940534:940997 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940534:940997 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:940534:940997 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940536:940996 [7] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940536:940996 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940536:940996 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:940536:940996 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940533:940993 [4] NCCL INFO Connected NVLS tree +ip-26-0-163-58:940533:940993 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940533:940993 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-58:940533:940993 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940536:940996 [7] NCCL INFO comm 0xaa8c2a0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x4251c4840ed1af9b - Init COMPLETE +ip-26-0-163-58:940534:940997 [5] NCCL INFO comm 0xc4fc660 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x4251c4840ed1af9b - Init COMPLETE +ip-26-0-163-58:940535:940995 [6] NCCL INFO comm 0xc2c1a00 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4251c4840ed1af9b - Init COMPLETE +ip-26-0-163-58:940533:940993 [4] NCCL INFO comm 0xa95e6d0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x4251c4840ed1af9b - Init COMPLETE +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967686:1968198 [0] NCCL INFO comm 0xaa2ef10 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb94acd845dcd0d04 - Init COMPLETE +ip-26-0-161-138:1967688:1968200 [2] NCCL INFO comm 0xc284a50 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb94acd845dcd0d04 - Init COMPLETE +ip-26-0-161-138:1967689:1968201 [3] NCCL INFO comm 0xb6516a0 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb94acd845dcd0d04 - Init COMPLETE +ip-26-0-161-138:1967687:1968199 [1] NCCL INFO comm 0xbd99ef0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb94acd845dcd0d04 - Init COMPLETE +ip-26-0-163-220:1685415:1685913 [1] NCCL INFO comm 0xb694c70 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x4c572e897855e7be - Init COMPLETE +ip-26-0-163-220:1685416:1685912 [2] NCCL INFO comm 0xb924060 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4c572e897855e7be - Init COMPLETE +ip-26-0-163-220:1685417:1685914 [3] NCCL INFO comm 0xa731c60 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x4c572e897855e7be - Init COMPLETE +ip-26-0-163-220:1685414:1685911 [0] NCCL INFO comm 0xc214a80 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4c572e897855e7be - Init COMPLETE +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO Connected NVLS tree +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO comm 0xc2ab3b0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5ec5cd891d1eb924 - Init START +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO comm 0xaefdfc0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5ec5cd891d1eb924 - Init START +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO comm 0xb67a830 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x945777dedb9c05b6 - Init START +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO comm 0xaa54590 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcdb1b77ae9d12a36 - Init START +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO comm 0xad43c30 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x945777dedb9c05b6 - Init START +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO comm 0xba1a880 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcdb1b77ae9d12a36 - Init START +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO comm 0xbdc2790 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x11ad393725ac26f3 - Init START +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO comm 0xb9d6f70 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x11ad393725ac26f3 - Init START +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685421:1685910 [7] NCCL INFO comm 0xc41f6a0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x56732c84596033a6 - Init COMPLETE +ip-26-0-163-220:1685420:1685908 [6] NCCL INFO comm 0xbc6fc70 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x56732c84596033a6 - Init COMPLETE +ip-26-0-163-220:1685419:1685907 [5] NCCL INFO comm 0xb008df0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x56732c84596033a6 - Init COMPLETE +ip-26-0-163-220:1685418:1685906 [4] NCCL INFO comm 0xb8db680 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x56732c84596033a6 - Init COMPLETE +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Connected all trees +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO NVLS comm 0xc1816c0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Connected all trees +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO NVLS comm 0xc288e30 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO NVLS comm 0xc475ae0 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Connected all trees +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO NVLS comm 0xc33dd00 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO Connected NVLS tree +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967693:1968204 [7] NCCL INFO comm 0xc288e30 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5363b24ac14b9b1 - Init COMPLETE +ip-26-0-161-138:1967691:1968205 [5] NCCL INFO comm 0xc475ae0 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5363b24ac14b9b1 - Init COMPLETE +ip-26-0-161-138:1967692:1968206 [6] NCCL INFO comm 0xc33dd00 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5363b24ac14b9b1 - Init COMPLETE +ip-26-0-161-138:1967690:1968203 [4] NCCL INFO comm 0xc1816c0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5363b24ac14b9b1 - Init COMPLETE +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO Connected all trees +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Connected all rings +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO Connected all trees +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Connected all rings +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO Connected all trees +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Connected all rings +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO Connected all trees +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967687:1968232 [1] NCCL INFO comm 0xbdc2790 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x11ad393725ac26f3 - Init COMPLETE +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO Connected all trees +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967689:1968228 [3] NCCL INFO comm 0xb67a830 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x945777dedb9c05b6 - Init COMPLETE +ip-26-0-161-138:1967686:1968231 [0] NCCL INFO comm 0xaa54590 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcdb1b77ae9d12a36 - Init COMPLETE +ip-26-0-161-138:1967688:1968230 [2] NCCL INFO comm 0xc2ab3b0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5ec5cd891d1eb924 - Init COMPLETE +12/28/2024 01:47:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [Optimizer Building] Using LearningRateForSP as learning rate +12/28/2024 01:47:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [ZeRO sharding] Size of optimizer params per rank: +12/28/2024 01:47:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [ZeRO sharding] DP Rank 0 has 117M out of 470M (25.00%) params' optimizer states +12/28/2024 01:47:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [ZeRO sharding] DP Rank 1 has 117M out of 470M (25.00%) params' optimizer states +12/28/2024 01:47:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [ZeRO sharding] DP Rank 2 has 117M out of 470M (25.00%) params' optimizer states +12/28/2024 01:47:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [ZeRO sharding] DP Rank 3 has 117M out of 470M (25.00%) params' optimizer states +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Connected all rings +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO Connected all trees +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Connected all rings +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO Connected all trees +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO Connected all trees +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207358:2207867 [1] NCCL INFO comm 0xb9d6f70 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x11ad393725ac26f3 - Init COMPLETE +ip-26-0-169-239:2207359:2207864 [2] NCCL INFO comm 0xaefdfc0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5ec5cd891d1eb924 - Init COMPLETE +ip-26-0-169-239:2207360:2207865 [3] NCCL INFO comm 0xad43c30 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x945777dedb9c05b6 - Init COMPLETE +ip-26-0-169-239:2207357:2207866 [0] NCCL INFO comm 0xba1a880 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcdb1b77ae9d12a36 - Init COMPLETE +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.token_position_embeddings | PP: 0/8 | Block rank: 0 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.0 | PP: 0/8 | Block rank: 1 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.1 | PP: 0/8 | Block rank: 2 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.2 | PP: 0/8 | Block rank: 3 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.3 | PP: 1/8 | Block rank: 0 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.4 | PP: 1/8 | Block rank: 1 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.5 | PP: 1/8 | Block rank: 2 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.6 | PP: 2/8 | Block rank: 0 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.7 | PP: 2/8 | Block rank: 1 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.8 | PP: 3/8 | Block rank: 0 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.9 | PP: 3/8 | Block rank: 1 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.10 | PP: 3/8 | Block rank: 2 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.11 | PP: 4/8 | Block rank: 0 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.12 | PP: 4/8 | Block rank: 1 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.13 | PP: 5/8 | Block rank: 0 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.14 | PP: 5/8 | Block rank: 1 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.15 | PP: 5/8 | Block rank: 2 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.final_layer_norm | PP: 6/8 | Block rank: 0 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.lm_head | PP: 6/8 | Block rank: 1 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.cast_to_fp32 | PP: 7/8 | Block rank: 0 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: loss | PP: 7/8 | Block rank: 1 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Using dummy data generator +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [Training Plan] There are 1 training stages +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [Stage Stable Training Stage] start from step 1 +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [Start training] datetime: 2024-12-28 01:47:34.072835 | mbs: 2 | grad_accum: 32 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/28/2024 01:47:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 3136.09MiB. Peak allocated 5440.00MiB. Peak reserved: 16452.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO comm 0xc5fb2b0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf019b61c3b5e4fdf - Init START +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO comm 0x19e6bbb0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf019b61c3b5e4fdf - Init START +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO comm 0x1a59a860 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2b88a4405145733b - Init START +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO comm 0xc65c500 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2b88a4405145733b - Init START +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO comm 0xc3e0480 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x718ae5960aafd72c - Init START +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO comm 0x17558ae0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x718ae5960aafd72c - Init START +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Using network Libfabric +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO comm 0xc543240 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x716fb56e9d15de89 - Init START +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO comm 0x1868c220 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x716fb56e9d15de89 - Init START +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 04/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 05/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 06/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 07/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 08/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 09/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 10/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 11/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 12/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 13/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 14/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 15/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 16/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 17/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 18/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 19/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 20/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 21/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 22/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Channel 23/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 04/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 05/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 06/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 07/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 08/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 09/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 10/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 11/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 04/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 12/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 05/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 06/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 07/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 13/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 08/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 14/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 09/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 15/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 10/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 11/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 16/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 17/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 12/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 18/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 13/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 19/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 14/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 20/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 15/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 21/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 16/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 22/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 17/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 18/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Channel 23/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Connected all rings +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO Connected all trees +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 19/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 20/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 21/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Connected all rings +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Connected all rings +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO Connected all trees +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 22/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967687:1968263 [1] NCCL INFO comm 0x1a59a860 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2b88a4405145733b - Init COMPLETE +ip-26-0-161-138:1967691:1968264 [5] NCCL INFO comm 0xc65c500 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x2b88a4405145733b - Init COMPLETE +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 00/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Channel 23/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 01/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Connected all rings +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO Connected all trees +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967692:1968261 [6] NCCL INFO comm 0xc5fb2b0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf019b61c3b5e4fdf - Init COMPLETE +ip-26-0-161-138:1967688:1968260 [2] NCCL INFO comm 0x19e6bbb0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf019b61c3b5e4fdf - Init COMPLETE +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 00/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 04/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 01/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 05/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 06/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 07/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 04/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 08/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 05/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 09/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 06/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 10/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 07/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 11/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 08/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 12/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 09/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 13/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 10/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 14/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 11/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 15/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 12/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 16/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 13/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 17/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 14/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 18/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 15/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 19/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 16/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Connected all rings +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO Connected all trees +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 20/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 17/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 21/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 18/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 22/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 19/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 23/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 20/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 21/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 24/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Connected all rings +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO Connected all trees +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 22/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 25/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 23/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 26/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 24/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 27/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Connected all rings +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO Connected all trees +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 25/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 28/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 26/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967693:1968270 [7] NCCL INFO comm 0xc543240 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x716fb56e9d15de89 - Init COMPLETE +ip-26-0-161-138:1967689:1968269 [3] NCCL INFO comm 0x1868c220 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x716fb56e9d15de89 - Init COMPLETE +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 00/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Connected all rings +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO Connected all trees +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 29/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 27/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 01/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 30/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 28/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967687:1968286 [1] NCCL INFO Channel 31/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 29/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 30/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 04/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967688:1968289 [2] NCCL INFO Channel 31/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 05/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 06/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968266 [0] NCCL INFO comm 0x17558ae0 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x718ae5960aafd72c - Init COMPLETE +ip-26-0-161-138:1967690:1968267 [4] NCCL INFO comm 0xc3e0480 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x718ae5960aafd72c - Init COMPLETE +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 00/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 07/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 01/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 08/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 09/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 04/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 10/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 11/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 05/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 12/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 06/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 13/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 14/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 15/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 07/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 16/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 08/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 17/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 09/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 18/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 10/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 19/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 11/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 20/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 12/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 13/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 21/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 14/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 15/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 22/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 16/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 23/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 17/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 24/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 18/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 25/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 19/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 26/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 20/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 27/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 21/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 28/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 22/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 29/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 23/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 30/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 24/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967689:1968294 [3] NCCL INFO Channel 31/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 25/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 26/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 27/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 28/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 29/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 30/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-161-138:1967686:1968296 [0] NCCL INFO Channel 31/1 : 0[0] -> 1[4] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO comm 0xb8d1a50 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x520fb0a9c0b39236 - Init START +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO comm 0x15085520 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x520fb0a9c0b39236 - Init START +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO comm 0xbaeccc0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xeecc81168b5db4e2 - Init START +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO comm 0x14ed5670 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xeecc81168b5db4e2 - Init START +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO comm 0xc456de0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc9704d2f6e211a32 - Init START +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO comm 0x15940ae0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc9704d2f6e211a32 - Init START +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO comm 0xa9a7be0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1bd113091827bf49 - Init START +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO comm 0x14aa70d0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1bd113091827bf49 - Init START +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Channel 00/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Channel 01/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Channel 02/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Channel 03/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 00/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 01/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 02/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 03/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Channel 00/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Channel 01/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Channel 02/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Channel 03/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 00/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 01/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 02/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 03/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Channel 00/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Channel 01/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Channel 02/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Channel 03/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 00/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 01/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 02/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 03/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Channel 00/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Channel 01/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Channel 02/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Channel 03/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 00/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 01/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 02/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 03/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Connected all rings +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO Connected all trees +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Connected all rings +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967691:1968310 [5] NCCL INFO comm 0x15085520 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x520fb0a9c0b39236 - Init COMPLETE +ip-26-0-161-138:1967691:1968325 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-138:1967691:1968325 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-163-220:1685415:1685947 [1] NCCL INFO comm 0xb8d1a50 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x520fb0a9c0b39236 - Init COMPLETE +ip-26-0-163-220:1685415:1685960 [1] NCCL INFO Channel 02/1 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-163-220:1685415:1685960 [1] NCCL INFO Channel 03/1 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Connected all rings +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO Connected all trees +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Connected all rings +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967692:1968312 [6] NCCL INFO comm 0x14ed5670 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xeecc81168b5db4e2 - Init COMPLETE +ip-26-0-161-138:1967692:1968326 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-138:1967692:1968326 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-163-220:1685416:1685948 [2] NCCL INFO comm 0xbaeccc0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xeecc81168b5db4e2 - Init COMPLETE +ip-26-0-163-220:1685416:1685961 [2] NCCL INFO Channel 02/1 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-163-220:1685416:1685961 [2] NCCL INFO Channel 03/1 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Connected all rings +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO Connected all trees +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967690:1968314 [4] NCCL INFO comm 0x15940ae0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc9704d2f6e211a32 - Init COMPLETE +ip-26-0-161-138:1967690:1968327 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-138:1967690:1968327 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685414:1685949 [0] NCCL INFO comm 0xc456de0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc9704d2f6e211a32 - Init COMPLETE +ip-26-0-163-220:1685414:1685962 [0] NCCL INFO Channel 02/1 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-163-220:1685414:1685962 [0] NCCL INFO Channel 03/1 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Connected all rings +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO Connected all trees +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-161-138:1967693:1968320 [7] NCCL INFO comm 0x14aa70d0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x1bd113091827bf49 - Init COMPLETE +ip-26-0-161-138:1967693:1968329 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-138:1967693:1968329 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-163-220:1685417:1685954 [3] NCCL INFO comm 0xa9a7be0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x1bd113091827bf49 - Init COMPLETE +ip-26-0-163-220:1685417:1685963 [3] NCCL INFO Channel 02/1 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-163-220:1685417:1685963 [3] NCCL INFO Channel 03/1 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO comm 0xb266030 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf7f4dd353c573619 - Init START +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO comm 0x143a3ff0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf7f4dd353c573619 - Init START +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO comm 0x145091d0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x27b54b10d7d28409 - Init START +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO comm 0xbecc680 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x27b54b10d7d28409 - Init START +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO comm 0x15ef8d20 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe52f5f2ef73284b3 - Init START +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO comm 0xbb20840 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe52f5f2ef73284b3 - Init START +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO comm 0x13650210 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x980b645efde219f8 - Init START +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO comm 0xc688bc0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x980b645efde219f8 - Init START +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Setting affinity for GPU 1 to 0fff,ffffffff +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Setting affinity for GPU 2 to 0fff,ffffffff +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Setting affinity for GPU 0 to 0fff,ffffffff +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Setting affinity for GPU 3 to 0fff,ffffffff +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 04/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 05/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 06/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 07/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 08/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 04/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 05/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 06/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 07/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 09/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 08/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 10/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 09/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 11/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 10/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 12/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 11/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 13/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 12/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 14/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 13/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 15/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 14/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 16/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 15/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 17/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 16/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 18/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 17/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 19/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 18/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 20/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 19/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 20/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 21/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 22/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 21/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Channel 23/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 22/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Channel 23/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 04/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 05/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 06/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 07/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 08/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 09/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 10/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 11/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 12/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 13/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 14/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 15/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 16/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 17/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 18/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 19/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 20/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Connected all rings +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO Connected all trees +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 21/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 22/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Channel 23/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685415:1685976 [1] NCCL INFO comm 0x143a3ff0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf7f4dd353c573619 - Init COMPLETE +ip-26-0-163-220:1685419:1685977 [5] NCCL INFO comm 0xb266030 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf7f4dd353c573619 - Init COMPLETE +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 00/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 01/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Connected all rings +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO Connected all trees +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 04/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 05/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Connected all rings +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 06/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Connected all rings +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO Connected all trees +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 07/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 08/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 09/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1685979 [2] NCCL INFO comm 0x145091d0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x27b54b10d7d28409 - Init COMPLETE +ip-26-0-163-220:1685420:1685980 [6] NCCL INFO comm 0xbecc680 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x27b54b10d7d28409 - Init COMPLETE +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 00/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 10/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 01/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 11/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 12/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1685982 [0] NCCL INFO comm 0x15ef8d20 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xe52f5f2ef73284b3 - Init COMPLETE +ip-26-0-163-220:1685418:1685983 [4] NCCL INFO comm 0xbb20840 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe52f5f2ef73284b3 - Init COMPLETE +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 00/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 04/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 13/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 05/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 14/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 01/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 06/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 15/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 07/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 16/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 08/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 17/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 04/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 09/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 18/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 10/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 05/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 06/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 11/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 07/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 12/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 08/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 13/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 19/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 14/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 09/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 20/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 15/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 10/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 21/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 16/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 11/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 22/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 17/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 12/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 23/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 18/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 13/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 24/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 14/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 19/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 25/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 20/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 15/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 26/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 21/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 16/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 27/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 17/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 22/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 28/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 18/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 23/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 29/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 19/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 24/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 30/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 25/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 20/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685415:1685999 [1] NCCL INFO Channel 31/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 21/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 26/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 27/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 22/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 28/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 23/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 29/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 24/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 30/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 25/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685416:1686006 [2] NCCL INFO Channel 31/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Connected all rings +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO Connected all trees +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 26/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Connected all rings +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 27/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 28/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 29/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 30/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685421:1685987 [7] NCCL INFO comm 0xc688bc0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x980b645efde219f8 - Init COMPLETE +ip-26-0-163-220:1685417:1685986 [3] NCCL INFO comm 0x13650210 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x980b645efde219f8 - Init COMPLETE +ip-26-0-163-220:1685414:1686007 [0] NCCL INFO Channel 31/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 00/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 01/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 04/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 05/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 06/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 07/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 08/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 09/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 10/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 11/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 12/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 13/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 14/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 15/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 16/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 17/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 18/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 19/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 20/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 21/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 22/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 23/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 24/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 25/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 26/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 27/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 28/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 29/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 30/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-220:1685417:1686012 [3] NCCL INFO Channel 31/1 : 0[3] -> 1[7] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:940530:941038 [1] NCCL INFO Using network Libfabric +ip-26-0-163-58:940530:941038 [1] NCCL INFO comm 0xb0da870 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xca6f4b640b5fa421 - Init START +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO comm 0x13bfb370 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xca6f4b640b5fa421 - Init START +ip-26-0-163-58:940530:941038 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940530:941038 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940530:941038 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-58:940530:941038 [1] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:940529:941041 [0] NCCL INFO Using network Libfabric +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO comm 0x141184b0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x1951755285d08e30 - Init START +ip-26-0-163-58:940529:941041 [0] NCCL INFO comm 0xaa92f50 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1951755285d08e30 - Init START +ip-26-0-163-58:940529:941041 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940530:941038 [1] NCCL INFO Channel 00/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:941038 [1] NCCL INFO Channel 01/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:941038 [1] NCCL INFO Channel 02/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:941038 [1] NCCL INFO Channel 03/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:941038 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:941038 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:941038 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940530:941038 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 00/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 01/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 02/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 03/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940529:941041 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,fff00000,00000000 +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-58:940529:941041 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-58:940529:941041 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Using network Libfabric +ip-26-0-163-58:940531:941046 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:940531:941046 [2] NCCL INFO comm 0xbd099c0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x455ec0ff2be0cc09 - Init START +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO comm 0x1433d270 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x455ec0ff2be0cc09 - Init START +ip-26-0-163-58:940531:941046 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940529:941041 [0] NCCL INFO Channel 00/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:941041 [0] NCCL INFO Channel 01/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:941041 [0] NCCL INFO Channel 02/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:941041 [0] NCCL INFO Channel 03/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:941041 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:941041 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:941041 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940529:941041 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 00/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 01/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 02/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 03/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,fff00000,00000000 +ip-26-0-163-58:940531:941046 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-58:940531:941046 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-58:940531:941046 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Using network Libfabric +ip-26-0-163-58:940532:941049 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:940532:941049 [3] NCCL INFO comm 0xaa59cc0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb3ac521d45be0505 - Init START +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO comm 0x154cacb0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb3ac521d45be0505 - Init START +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940532:941049 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940531:941046 [2] NCCL INFO Channel 00/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:941046 [2] NCCL INFO Channel 01/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:941046 [2] NCCL INFO Channel 02/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:941046 [2] NCCL INFO Channel 03/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:941046 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:941046 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:941046 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940531:941046 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 00/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 01/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 02/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 03/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,fff00000,00000000 +ip-26-0-163-58:940532:941049 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:940532:941049 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-163-58:940532:941049 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Connected all rings +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO Connected all trees +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940530:941038 [1] NCCL INFO Connected all rings +ip-26-0-163-58:940530:941038 [1] NCCL INFO Connected all trees +ip-26-0-163-58:940530:941038 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940530:941038 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940530:941038 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685419:1686027 [5] NCCL INFO comm 0x13bfb370 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xca6f4b640b5fa421 - Init COMPLETE +ip-26-0-163-220:1685419:1686042 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-163-220:1685419:1686042 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-163-58:940530:941038 [1] NCCL INFO comm 0xb0da870 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xca6f4b640b5fa421 - Init COMPLETE +ip-26-0-163-58:940530:941052 [1] NCCL INFO Channel 02/1 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-163-58:940530:941052 [1] NCCL INFO Channel 03/1 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Connected all rings +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO Connected all trees +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685418:1686031 [4] NCCL INFO comm 0x141184b0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x1951755285d08e30 - Init COMPLETE +ip-26-0-163-58:940532:941049 [3] NCCL INFO Channel 00/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:941049 [3] NCCL INFO Channel 01/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:941049 [3] NCCL INFO Channel 02/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:941049 [3] NCCL INFO Channel 03/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1686043 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-163-58:940532:941049 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685418:1686043 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-163-58:940532:941049 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:941049 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940532:941049 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 00/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 01/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 02/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 03/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940529:941041 [0] NCCL INFO Connected all rings +ip-26-0-163-58:940529:941041 [0] NCCL INFO Connected all trees +ip-26-0-163-58:940529:941041 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940529:941041 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940529:941041 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940529:941041 [0] NCCL INFO comm 0xaa92f50 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x1951755285d08e30 - Init COMPLETE +ip-26-0-163-58:940529:941053 [0] NCCL INFO Channel 02/1 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-163-58:940529:941053 [0] NCCL INFO Channel 03/1 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-163-58:940531:941046 [2] NCCL INFO Connected all rings +ip-26-0-163-58:940531:941046 [2] NCCL INFO Connected all trees +ip-26-0-163-58:940531:941046 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940531:941046 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940531:941046 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Connected all rings +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO Connected all trees +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940531:941046 [2] NCCL INFO comm 0xbd099c0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x455ec0ff2be0cc09 - Init COMPLETE +ip-26-0-163-58:940531:941054 [2] NCCL INFO Channel 02/1 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-163-58:940531:941054 [2] NCCL INFO Channel 03/1 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-163-220:1685420:1686035 [6] NCCL INFO comm 0x1433d270 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x455ec0ff2be0cc09 - Init COMPLETE +ip-26-0-163-220:1685420:1686044 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-163-220:1685420:1686044 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-163-58:940532:941049 [3] NCCL INFO Connected all rings +ip-26-0-163-58:940532:941049 [3] NCCL INFO Connected all trees +ip-26-0-163-58:940532:941049 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940532:941049 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940532:941049 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Connected all rings +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO Connected all trees +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940532:941049 [3] NCCL INFO comm 0xaa59cc0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb3ac521d45be0505 - Init COMPLETE +ip-26-0-163-220:1685421:1686039 [7] NCCL INFO comm 0x154cacb0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb3ac521d45be0505 - Init COMPLETE +ip-26-0-163-58:940532:941055 [3] NCCL INFO Channel 02/1 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-163-58:940532:941055 [3] NCCL INFO Channel 03/1 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-163-220:1685421:1686046 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-163-220:1685421:1686046 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-58:940534:941067 [5] NCCL INFO Using network Libfabric +ip-26-0-163-58:940530:941066 [1] NCCL INFO Using network Libfabric +ip-26-0-163-58:940534:941067 [5] NCCL INFO comm 0xc720f50 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x472665eed91a4a4c - Init START +ip-26-0-163-58:940530:941066 [1] NCCL INFO comm 0x13526f90 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x472665eed91a4a4c - Init START +ip-26-0-163-58:940534:941067 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940530:941066 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-58:940529:941069 [0] NCCL INFO Using network Libfabric +ip-26-0-163-58:940533:941070 [4] NCCL INFO Using network Libfabric +ip-26-0-163-58:940529:941069 [0] NCCL INFO comm 0x13447340 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x87b199a0ae8934d9 - Init START +ip-26-0-163-58:940533:941070 [4] NCCL INFO comm 0xabb8d30 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x87b199a0ae8934d9 - Init START +ip-26-0-163-58:940533:941070 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940529:941069 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940534:941067 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-58:940531:941073 [2] NCCL INFO Using network Libfabric +ip-26-0-163-58:940535:941074 [6] NCCL INFO Using network Libfabric +ip-26-0-163-58:940535:941074 [6] NCCL INFO comm 0xc51bfe0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4e9d2581ce32ee4b - Init START +ip-26-0-163-58:940531:941073 [2] NCCL INFO comm 0x1475aaa0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4e9d2581ce32ee4b - Init START +ip-26-0-163-58:940535:941074 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940531:941073 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-163-58:940534:941067 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-163-58:940534:941067 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-163-58:940530:941066 [1] NCCL INFO P2P Chunksize set to 524288 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-58:940532:941078 [3] NCCL INFO Using network Libfabric +ip-26-0-163-58:940536:941079 [7] NCCL INFO Using network Libfabric +ip-26-0-163-58:940536:941079 [7] NCCL INFO comm 0xaccff60 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7584bbd4e5b9a05f - Init START +ip-26-0-163-58:940532:941078 [3] NCCL INFO comm 0x12f0a1a0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7584bbd4e5b9a05f - Init START +ip-26-0-163-58:940536:941079 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940532:941078 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-163-58:940533:941070 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-163-58:940533:941070 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-163-58:940529:941069 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 04/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-163-58:940535:941074 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-163-58:940535:941074 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-163-58:940531:941073 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 05/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 06/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 07/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 08/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 09/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 10/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 11/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 12/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 13/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 14/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-163-58:940536:941079 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-163-58:940536:941079 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-163-58:940532:941078 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 15/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 16/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 17/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 18/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 19/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 20/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 21/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 22/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Channel 23/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941067 [5] NCCL INFO Connected all rings +ip-26-0-163-58:940534:941067 [5] NCCL INFO Connected all trees +ip-26-0-163-58:940534:941067 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940534:941067 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940534:941067 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940530:941066 [1] NCCL INFO Connected all rings +ip-26-0-163-58:940530:941066 [1] NCCL INFO Connected all trees +ip-26-0-163-58:940530:941066 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940530:941066 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940530:941066 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940530:941066 [1] NCCL INFO comm 0x13526f90 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x472665eed91a4a4c - Init COMPLETE +ip-26-0-163-58:940534:941067 [5] NCCL INFO comm 0xc720f50 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x472665eed91a4a4c - Init COMPLETE +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 00/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 01/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 04/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 05/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 06/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 04/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 05/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 06/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 07/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 07/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 08/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 08/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 09/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 09/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 10/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 10/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 11/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 11/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 12/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 12/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 13/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 13/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 14/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 14/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 15/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 15/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 16/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 16/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 17/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 17/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 04/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 18/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 18/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 05/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 19/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 19/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 06/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 20/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 20/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 07/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 21/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 21/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 08/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 22/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 22/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 09/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 23/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO Connected all rings +ip-26-0-163-58:940533:941070 [4] NCCL INFO Connected all trees +ip-26-0-163-58:940533:941070 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940533:941070 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940533:941070 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940531:941073 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 10/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Channel 23/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 24/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 25/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 11/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 12/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 13/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940529:941069 [0] NCCL INFO Connected all rings +ip-26-0-163-58:940529:941069 [0] NCCL INFO Connected all trees +ip-26-0-163-58:940529:941069 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940529:941069 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940529:941069 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 26/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 14/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 27/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 15/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 28/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 16/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 29/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940533:941070 [4] NCCL INFO comm 0xabb8d30 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x87b199a0ae8934d9 - Init COMPLETE +ip-26-0-163-58:940529:941069 [0] NCCL INFO comm 0x13447340 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x87b199a0ae8934d9 - Init COMPLETE +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 00/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 17/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 30/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 01/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 18/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940530:941091 [1] NCCL INFO Channel 31/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 19/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 20/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 04/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 21/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 05/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 22/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 06/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO Channel 23/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 07/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 08/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 09/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 10/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 11/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 12/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 13/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 14/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 15/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 16/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 17/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 18/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 19/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 20/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO Connected all rings +ip-26-0-163-58:940535:941074 [6] NCCL INFO Connected all trees +ip-26-0-163-58:940535:941074 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940535:941074 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940535:941074 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 21/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 22/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 23/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 24/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940531:941073 [2] NCCL INFO Connected all rings +ip-26-0-163-58:940531:941073 [2] NCCL INFO Connected all trees +ip-26-0-163-58:940531:941073 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940531:941073 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940531:941073 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 25/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 26/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 27/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940535:941074 [6] NCCL INFO comm 0xc51bfe0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x4e9d2581ce32ee4b - Init COMPLETE +ip-26-0-163-58:940531:941073 [2] NCCL INFO comm 0x1475aaa0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x4e9d2581ce32ee4b - Init COMPLETE +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 00/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 01/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 28/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 29/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 30/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 04/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940529:941094 [0] NCCL INFO Channel 31/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 05/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 06/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 07/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 08/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 09/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 10/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO Connected all rings +ip-26-0-163-58:940536:941079 [7] NCCL INFO Connected all trees +ip-26-0-163-58:940536:941079 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940536:941079 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940536:941079 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940532:941078 [3] NCCL INFO Connected all rings +ip-26-0-163-58:940532:941078 [3] NCCL INFO Connected all trees +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 11/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941078 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940532:941078 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940532:941078 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 12/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 13/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 14/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940536:941079 [7] NCCL INFO comm 0xaccff60 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7584bbd4e5b9a05f - Init COMPLETE +ip-26-0-163-58:940532:941078 [3] NCCL INFO comm 0x12f0a1a0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7584bbd4e5b9a05f - Init COMPLETE +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 00/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 15/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 16/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 01/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 17/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 18/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 19/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 04/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 20/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 05/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 21/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 06/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 22/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 07/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 23/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 08/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 24/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 09/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 25/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 10/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 26/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 11/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 12/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 27/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 13/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 28/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 14/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 29/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 15/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 30/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 16/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940531:941098 [2] NCCL INFO Channel 31/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 17/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 18/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 19/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 20/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 21/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 22/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 23/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 24/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 25/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 26/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 27/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 28/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 29/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 30/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-163-58:940532:941103 [3] NCCL INFO Channel 31/1 : 0[3] -> 1[7] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-58:940534:941115 [5] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO comm 0xc54a8a0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6abdc7c5c5472d83 - Init START +ip-26-0-163-58:940534:941115 [5] NCCL INFO comm 0x160a1660 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6abdc7c5c5472d83 - Init START +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940534:941115 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-163-58:940534:941115 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-58:940534:941115 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:940534:941115 [5] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-58:940533:941118 [4] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO comm 0xc4b85a0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9ca82ca0219ee4d6 - Init START +ip-26-0-163-58:940533:941118 [4] NCCL INFO comm 0x12edc780 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9ca82ca0219ee4d6 - Init START +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940533:941118 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Channel 00/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Channel 01/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Channel 02/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Channel 03/0 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 00/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 01/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 02/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 03/0 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940534:941115 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-163-58:940533:941118 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-58:940533:941118 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:940533:941118 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Channel 00/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Channel 01/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Channel 02/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Channel 03/0 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 00/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 01/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 02/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 03/0 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-163-58:940533:941118 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-58:940535:941123 [6] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO comm 0xba729d0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb647ca739fbbea8d - Init START +ip-26-0-163-58:940535:941123 [6] NCCL INFO comm 0x1646bc20 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb647ca739fbbea8d - Init START +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940535:941123 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940535:941123 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-163-58:940535:941123 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:940535:941123 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940534:941115 [5] NCCL INFO Connected all rings +ip-26-0-163-58:940534:941115 [5] NCCL INFO Connected all trees +ip-26-0-163-58:940534:941115 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940534:941115 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940534:941115 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940534:941115 [5] NCCL INFO comm 0x160a1660 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x6abdc7c5c5472d83 - Init COMPLETE +ip-26-0-163-58:940534:941126 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-163-58:940534:941126 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[1] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO Connected all trees +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207358:2207901 [1] NCCL INFO comm 0xc54a8a0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x6abdc7c5c5472d83 - Init COMPLETE +ip-26-0-169-239:2207358:2207910 [1] NCCL INFO Channel 02/1 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-169-239:2207358:2207910 [1] NCCL INFO Channel 03/1 : 0[5] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-163-58:940533:941118 [4] NCCL INFO Connected all rings +ip-26-0-163-58:940533:941118 [4] NCCL INFO Connected all trees +ip-26-0-163-58:940533:941118 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940533:941118 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940533:941118 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Channel 00/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Channel 01/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Channel 02/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Channel 03/0 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 00/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 01/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 02/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 03/0 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940535:941123 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-163-58:940533:941118 [4] NCCL INFO comm 0x12edc780 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x9ca82ca0219ee4d6 - Init COMPLETE +ip-26-0-163-58:940533:941127 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-163-58:940533:941127 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[0] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO Connected all trees +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207357:2207904 [0] NCCL INFO comm 0xc4b85a0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x9ca82ca0219ee4d6 - Init COMPLETE +ip-26-0-169-239:2207357:2207911 [0] NCCL INFO Channel 02/1 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-169-239:2207357:2207911 [0] NCCL INFO Channel 03/1 : 0[4] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-163-58:940536:941129 [7] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO comm 0xb7dd650 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf3e5b4a4041da526 - Init START +ip-26-0-163-58:940536:941129 [7] NCCL INFO comm 0x13141ac0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf3e5b4a4041da526 - Init START +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940536:941129 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-163-58:940536:941129 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940536:941129 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-163-58:940536:941129 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-163-58:940535:941123 [6] NCCL INFO Connected all rings +ip-26-0-163-58:940535:941123 [6] NCCL INFO Connected all trees +ip-26-0-163-58:940535:941123 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940535:941123 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940535:941123 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940535:941123 [6] NCCL INFO comm 0x1646bc20 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb647ca739fbbea8d - Init COMPLETE +ip-26-0-163-58:940535:941132 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Connected all rings +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO Connected all trees +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940535:941132 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[2] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Channel 00/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Channel 01/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Channel 02/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Channel 03/0 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 00/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 01/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 02/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 03/0 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-163-58:940536:941129 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-169-239:2207359:2207907 [2] NCCL INFO comm 0xba729d0 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb647ca739fbbea8d - Init COMPLETE +ip-26-0-169-239:2207359:2207915 [2] NCCL INFO Channel 02/1 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-169-239:2207359:2207915 [2] NCCL INFO Channel 03/1 : 0[6] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-163-58:940536:941129 [7] NCCL INFO Connected all rings +ip-26-0-163-58:940536:941129 [7] NCCL INFO Connected all trees +ip-26-0-163-58:940536:941129 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-163-58:940536:941129 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-163-58:940536:941129 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Connected all rings +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO Connected all trees +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-163-58:940536:941129 [7] NCCL INFO comm 0x13141ac0 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf3e5b4a4041da526 - Init COMPLETE +ip-26-0-163-58:940536:941135 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-163-58:940536:941135 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[3] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-169-239:2207360:2207912 [3] NCCL INFO comm 0xb7dd650 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf3e5b4a4041da526 - Init COMPLETE +ip-26-0-169-239:2207360:2207917 [3] NCCL INFO Channel 02/1 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-169-239:2207360:2207917 [3] NCCL INFO Channel 03/1 : 0[7] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO comm 0xdc15220 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x54f63ae47dbd20ca - Init START +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO comm 0xaccde00 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x54f63ae47dbd20ca - Init START +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO comm 0xdbd0080 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7ea30c91f458b273 - Init START +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO comm 0xad95ac0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7ea30c91f458b273 - Init START +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO comm 0xd0f7590 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd6eb5f1b360b485 - Init START +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO comm 0xa22c4f0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd6eb5f1b360b485 - Init START +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO P2P Chunksize set to 524288 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO comm 0xcf3b1a0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6cf255399bedf228 - Init START +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO comm 0xb2e6cd0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6cf255399bedf228 - Init START +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 04/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 05/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 06/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 07/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 08/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 04/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 09/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 05/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 10/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 06/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 11/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 07/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 12/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 13/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 08/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 09/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 10/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 14/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 15/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 11/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 16/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 12/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 17/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 13/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 18/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 14/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 19/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 15/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 20/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 16/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 21/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 17/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 22/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 18/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Channel 23/0 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 19/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 20/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 21/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 22/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Channel 23/0 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Connected all rings +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO Connected all trees +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Connected all rings +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO Connected all trees +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 00/24 : 0 1 +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] -1/-1/-1->1->0 [3] -1/-1/-1->1->0 [4] -1/-1/-1->1->0 [5] -1/-1/-1->1->0 [6] 0/-1/-1->1->-1 [7] 0/-1/-1->1->-1 [8] 0/-1/-1->1->-1 [9] 0/-1/-1->1->-1 [10] 0/-1/-1->1->-1 [11] 0/-1/-1->1->-1 [12] -1/-1/-1->1->0 [13] -1/-1/-1->1->0 [14] -1/-1/-1->1->0 [15] -1/-1/-1->1->0 [16] -1/-1/-1->1->0 [17] -1/-1/-1->1->0 [18] 0/-1/-1->1->-1 [19] 0/-1/-1->1->-1 [20] 0/-1/-1->1->-1 [21] 0/-1/-1->1->-1 [22] 0/-1/-1->1->-1 [23] 0/-1/-1->1->-1 +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 01/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 02/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 03/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 04/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 05/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 06/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 07/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 08/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 09/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 10/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 11/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 12/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 13/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 14/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 15/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 16/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 17/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 18/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 19/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 20/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 21/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 22/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 23/24 : 0 1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] -1/-1/-1->0->1 [7] -1/-1/-1->0->1 [8] -1/-1/-1->0->1 [9] -1/-1/-1->0->1 [10] -1/-1/-1->0->1 [11] -1/-1/-1->0->1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] -1/-1/-1->0->1 [19] -1/-1/-1->0->1 [20] -1/-1/-1->0->1 [21] -1/-1/-1->0->1 [22] -1/-1/-1->0->1 [23] -1/-1/-1->0->1 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 04/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Connected all rings +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO Connected all trees +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 05/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 06/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 07/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 08/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Connected all rings +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO Connected all trees +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 09/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 04/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 10/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 05/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 11/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 12/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 06/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 13/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 07/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 14/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 08/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 15/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 09/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 16/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 10/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 17/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 11/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 18/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207357:2207927 [0] NCCL INFO comm 0xdc15220 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x54f63ae47dbd20ca - Init COMPLETE +ip-26-0-169-239:2207361:2207928 [4] NCCL INFO comm 0xaccde00 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x54f63ae47dbd20ca - Init COMPLETE +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 12/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 00/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 19/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 13/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 20/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 01/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 14/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 21/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 22/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 15/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Channel 23/0 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 16/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 04/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 17/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 05/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207362:2207931 [5] NCCL INFO comm 0xad95ac0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7ea30c91f458b273 - Init COMPLETE +ip-26-0-169-239:2207358:2207930 [1] NCCL INFO comm 0xdbd0080 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7ea30c91f458b273 - Init COMPLETE +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 00/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 18/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 06/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 19/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 01/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 07/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 20/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 08/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 21/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 09/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 22/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 04/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 10/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Channel 23/0 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 05/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 11/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 06/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 12/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 07/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 13/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 08/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 14/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 09/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 15/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 10/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 16/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 04/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 17/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 05/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 18/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 06/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 19/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 07/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 20/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 08/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 21/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 11/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 09/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 22/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 04/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 10/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 12/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 23/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 05/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 13/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 06/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 14/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 07/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 24/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 08/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 25/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 11/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 09/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 15/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 26/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 12/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 10/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 16/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 27/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 13/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 11/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 17/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 28/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 14/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 12/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 15/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 18/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 16/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 19/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 13/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 29/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 17/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 20/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 14/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 30/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 18/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 21/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 15/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207357:2207954 [0] NCCL INFO Channel 31/1 : 0[0] -> 1[4] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 19/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 22/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 16/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 20/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 23/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 17/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 21/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 24/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 18/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 22/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 25/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 19/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Channel 23/0 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 26/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 20/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 27/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 21/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 28/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 22/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 29/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Channel 23/0 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 30/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Connected all rings +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO Connected all trees +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207358:2207955 [1] NCCL INFO Channel 31/1 : 0[1] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Connected all rings +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO Connected all trees +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207363:2207936 [6] NCCL INFO comm 0xa22c4f0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xfd6eb5f1b360b485 - Init COMPLETE +ip-26-0-169-239:2207359:2207935 [2] NCCL INFO comm 0xd0f7590 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xfd6eb5f1b360b485 - Init COMPLETE +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 00/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 01/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 04/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 05/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 06/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 07/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 08/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 09/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 10/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 11/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 12/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 13/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 14/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 15/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 16/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 17/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 18/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 19/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 20/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 21/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 22/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 23/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 24/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Connected all rings +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO Connected all trees +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Connected all rings +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO Connected all trees +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO 24 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 25/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 26/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 27/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 28/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 29/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 30/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207359:2207961 [2] NCCL INFO Channel 31/1 : 0[2] -> 1[6] via P2P/IPC +ip-26-0-169-239:2207364:2207943 [7] NCCL INFO comm 0xb2e6cd0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x6cf255399bedf228 - Init COMPLETE +ip-26-0-169-239:2207360:2207942 [3] NCCL INFO comm 0xcf3b1a0 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x6cf255399bedf228 - Init COMPLETE +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 00/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 01/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 04/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 05/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 06/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 07/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 08/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 09/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 10/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 11/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 12/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 13/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 14/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 15/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 16/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 17/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 18/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 19/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 20/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 21/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 22/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 23/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 24/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 25/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 26/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 27/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 28/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 29/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 30/1 : 0[3] -> 1[7] via P2P/IPC +ip-26-0-169-239:2207360:2207965 [3] NCCL INFO Channel 31/1 : 0[3] -> 1[7] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 00/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 01/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 04/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 05/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 06/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 07/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 08/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 09/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 10/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 11/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 12/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 13/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 14/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 15/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 16/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 17/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 18/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 19/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 20/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 21/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 22/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 23/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 24/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 25/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 26/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 27/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 28/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 29/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 30/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-169-239:2207361:2208009 [4] NCCL INFO Channel 31/1 : 1[4] -> 0[0] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 00/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 01/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 04/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 05/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 06/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 07/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 08/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 09/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 10/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 11/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 12/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 13/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 14/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 15/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 16/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 17/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 18/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 19/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 20/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 21/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 22/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 23/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 24/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 25/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 26/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 27/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 28/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 29/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 30/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-169-239:2207362:2208027 [5] NCCL INFO Channel 31/1 : 1[5] -> 0[1] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 00/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 01/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 04/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 05/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 06/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 07/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 08/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 09/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 10/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 11/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 12/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 13/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 14/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 15/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 16/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 17/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 18/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 19/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 20/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 21/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 22/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 23/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 24/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 25/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 26/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 27/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 28/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 29/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 30/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-169-239:2207363:2208036 [6] NCCL INFO Channel 31/1 : 1[6] -> 0[2] via P2P/IPC +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 00/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 01/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 04/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 05/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 06/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 07/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 08/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 09/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 10/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 11/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 12/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 13/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 14/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 15/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 16/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 17/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 18/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 19/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 20/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 21/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 22/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 23/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 24/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 25/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 26/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 27/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 28/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 29/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 30/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-169-239:2207364:2208038 [7] NCCL INFO Channel 31/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940533:941210 [4] NCCL INFO Channel 02/1 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-163-58:940533:941210 [4] NCCL INFO Channel 03/1 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-163-58:940534:941211 [5] NCCL INFO Channel 02/1 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-163-58:940534:941211 [5] NCCL INFO Channel 03/1 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-163-58:940535:941212 [6] NCCL INFO Channel 02/1 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-163-58:940535:941212 [6] NCCL INFO Channel 03/1 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-169-239:2207357:2208042 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-169-239:2207357:2208042 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-169-239:2207358:2208043 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-169-239:2207358:2208043 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-163-58:940536:941213 [7] NCCL INFO Channel 02/1 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-163-58:940536:941213 [7] NCCL INFO Channel 03/1 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-169-239:2207359:2208044 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-169-239:2207359:2208044 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-169-239:2207360:2208045 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-169-239:2207360:2208045 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 00/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 00/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 01/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 01/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 04/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 04/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 05/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 05/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 06/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 06/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 07/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 07/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 08/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 08/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 09/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 09/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 10/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 10/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 11/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 11/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 12/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 12/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 13/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 13/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 14/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 14/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 15/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 15/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 16/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 17/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 16/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 18/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 17/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 19/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 18/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 19/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 20/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 20/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 21/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 21/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 22/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 22/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 23/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 23/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 24/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 24/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 25/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 25/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 26/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 26/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 27/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 27/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 28/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 28/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 29/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 29/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 30/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 30/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940534:941216 [5] NCCL INFO Channel 31/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-58:940533:941217 [4] NCCL INFO Channel 31/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 00/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 01/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 04/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 05/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 06/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 07/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 08/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 09/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 10/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 11/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 12/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 13/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 14/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 15/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 16/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 17/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 18/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 19/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 20/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 21/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 22/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 23/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 24/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 25/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 26/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 27/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 28/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 29/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 30/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940535:941218 [6] NCCL INFO Channel 31/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 00/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 01/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 04/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 05/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 06/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 07/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 08/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 09/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 10/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 11/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 12/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 13/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 14/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 15/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 16/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 17/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 18/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 19/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 20/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 21/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 22/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 23/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 24/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 25/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 26/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 27/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 28/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 29/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 30/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-58:940536:941219 [7] NCCL INFO Channel 31/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685418:1686137 [4] NCCL INFO Channel 02/1 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-163-220:1685418:1686137 [4] NCCL INFO Channel 03/1 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-163-220:1685419:1686138 [5] NCCL INFO Channel 02/1 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-163-220:1685419:1686138 [5] NCCL INFO Channel 03/1 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-163-220:1685421:1686139 [7] NCCL INFO Channel 02/1 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-163-220:1685421:1686139 [7] NCCL INFO Channel 03/1 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-163-58:940529:941221 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-163-58:940529:941221 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-163-220:1685420:1686140 [6] NCCL INFO Channel 02/1 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-163-220:1685420:1686140 [6] NCCL INFO Channel 03/1 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-163-58:940530:941222 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-163-58:940530:941222 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-163-58:940531:941224 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-163-58:940531:941224 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-163-58:940532:941225 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-163-58:940532:941225 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 00/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 01/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 04/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 05/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 06/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 07/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 08/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 09/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 10/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 11/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 12/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 13/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 14/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 15/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 16/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 17/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 18/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 19/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 20/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 21/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 22/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 23/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 24/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 25/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 26/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 27/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 28/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 29/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 30/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685418:1686142 [4] NCCL INFO Channel 31/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 00/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 01/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 00/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 04/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 01/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 05/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 06/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 07/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 04/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 08/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 05/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 09/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 06/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 00/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 10/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 07/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 11/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 01/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 08/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 12/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 09/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 13/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 10/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 14/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 04/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 11/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 15/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 05/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 12/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 16/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 06/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 13/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 17/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 07/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 14/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 18/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 15/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 08/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 19/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 16/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 09/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 20/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 17/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 10/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 21/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 18/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 11/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 19/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 22/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 12/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 23/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 20/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 13/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 21/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 24/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 14/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 25/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 22/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 15/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 26/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 23/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 16/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 27/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 24/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 17/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 28/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 25/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 18/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 29/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 26/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 19/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 30/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 27/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 20/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685420:1686143 [6] NCCL INFO Channel 31/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 28/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 21/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 29/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 22/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 30/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 23/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685419:1686144 [5] NCCL INFO Channel 31/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 24/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 25/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 26/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 27/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 28/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 29/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 30/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-163-220:1685421:1686145 [7] NCCL INFO Channel 31/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967690:1968435 [4] NCCL INFO Channel 02/1 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-161-138:1967690:1968435 [4] NCCL INFO Channel 03/1 : 1[0] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-163-220:1685414:1686162 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-163-220:1685414:1686162 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[4] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-161-138:1967692:1968448 [6] NCCL INFO Channel 02/1 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-138:1967692:1968448 [6] NCCL INFO Channel 03/1 : 1[2] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-161-138:1967693:1968449 [7] NCCL INFO Channel 02/1 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-138:1967693:1968449 [7] NCCL INFO Channel 03/1 : 1[3] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-161-138:1967691:1968450 [5] NCCL INFO Channel 02/1 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-161-138:1967691:1968450 [5] NCCL INFO Channel 03/1 : 1[1] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-163-220:1685416:1686163 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-163-220:1685416:1686163 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[6] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-163-220:1685415:1686164 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-163-220:1685415:1686164 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[5] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-163-220:1685417:1686165 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-163-220:1685417:1686165 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[7] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 00/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 01/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 04/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 05/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 06/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 07/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 08/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 09/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 10/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 11/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 12/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 13/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 14/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 15/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 16/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 17/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 18/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 19/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 20/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 21/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 22/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 23/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 24/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 25/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 26/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 27/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 28/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 29/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 30/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967690:1968452 [4] NCCL INFO Channel 31/1 : 1[4] -> 0[0] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 00/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 01/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 04/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 05/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 06/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 07/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 08/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 09/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 10/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 11/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 12/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 13/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 14/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 15/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 16/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 17/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 18/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 19/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 20/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 21/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 22/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 23/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 24/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 25/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 26/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 27/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 28/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 29/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 30/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967692:1968453 [6] NCCL INFO Channel 31/1 : 1[6] -> 0[2] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 00/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 01/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 04/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 00/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 05/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 01/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 06/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 07/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 08/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 04/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 09/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 05/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 10/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 06/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 11/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 07/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 12/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 08/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 13/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 09/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 14/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 10/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 15/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 11/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 16/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 12/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 17/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 13/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 18/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 14/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 19/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 15/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 20/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 16/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 21/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 17/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 22/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 18/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 23/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 19/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 24/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 20/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 25/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 21/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 26/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 22/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 27/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 23/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 28/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 24/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 29/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 25/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 30/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 26/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967693:1968455 [7] NCCL INFO Channel 31/1 : 1[7] -> 0[3] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 27/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 28/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 29/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 30/1 : 1[5] -> 0[1] via P2P/IPC +ip-26-0-161-138:1967691:1968456 [5] NCCL INFO Channel 31/1 : 1[5] -> 0[1] via P2P/IPC +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Using network Libfabric +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO comm 0x19cf06f0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe68e38d0768d78fc - Init START +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO comm 0x1a308dc0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe68e38d0768d78fc - Init START +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO comm 0x19250500 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe68e38d0768d78fc - Init START +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO comm 0x19db9170 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe68e38d0768d78fc - Init START +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 00/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 01/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 02/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 03/24 : 0 1 2 3 +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 04/24 : 0 1 2 3 +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] -1/-1/-1->3->2 [3] -1/-1/-1->3->2 [4] -1/-1/-1->3->2 [5] -1/-1/-1->3->2 [6] -1/-1/-1->3->2 [7] -1/-1/-1->3->2 [8] -1/-1/-1->3->2 [9] -1/-1/-1->3->2 [10] -1/-1/-1->3->2 [11] -1/-1/-1->3->2 [12] -1/-1/-1->3->2 [13] -1/-1/-1->3->2 [14] -1/-1/-1->3->2 [15] -1/-1/-1->3->2 [16] -1/-1/-1->3->2 [17] -1/-1/-1->3->2 [18] -1/-1/-1->3->2 [19] -1/-1/-1->3->2 [20] -1/-1/-1->3->2 [21] -1/-1/-1->3->2 [22] -1/-1/-1->3->2 [23] -1/-1/-1->3->2 +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 05/24 : 0 1 2 3 +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 06/24 : 0 1 2 3 +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 07/24 : 0 1 2 3 +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 08/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 09/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 10/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 11/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 12/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 13/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 14/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 15/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 16/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 17/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 18/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 19/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 20/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 21/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 22/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 23/24 : 0 1 2 3 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 04/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 04/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 04/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 04/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 05/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 05/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 05/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 05/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 06/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 06/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 06/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 06/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 07/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 07/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 07/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 07/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 08/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 08/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 08/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 08/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 09/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 09/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 09/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 09/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 10/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 10/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 10/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 10/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 11/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 11/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 11/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 11/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 12/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 12/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 12/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 12/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 13/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 13/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 13/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 13/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 14/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 14/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 14/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 14/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 15/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 15/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 15/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 15/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 16/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 16/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 16/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 16/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 17/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 17/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 17/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 17/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 18/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 18/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 18/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 18/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 19/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 19/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 19/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 19/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 20/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 20/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 20/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 20/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 21/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 21/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 21/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 21/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 22/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 22/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 22/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 22/0 : 0[4] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 23/0 : 3[7] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 23/0 : 2[6] -> 3[7] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 23/0 : 1[5] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Channel 23/0 : 0[4] -> 1[5] via P2P/IPC +12/28/2024 01:48:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 3201.66MiB. Peak allocated 23515.65MiB. Peak reserved: 24192.00MiB +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Connected all rings +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Connected all rings +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 02/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Connected all rings +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 03/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Connected all rings +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 04/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 05/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 06/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 07/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 08/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 09/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 10/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 11/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 12/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 13/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 14/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 15/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 16/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 17/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 18/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 19/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 20/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 21/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 22/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Channel 23/0 : 3[7] -> 2[6] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 04/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 05/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 04/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 06/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 05/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 07/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 06/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 08/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 07/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 09/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 08/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 10/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 09/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 11/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 10/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 12/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 11/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 13/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 12/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 14/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 13/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 15/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 14/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 16/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 15/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 17/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 16/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 18/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 17/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 19/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 18/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 20/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 19/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 21/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 20/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 22/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 21/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Channel 23/0 : 2[6] -> 1[5] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 22/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Channel 23/0 : 1[5] -> 0[4] via P2P/IPC +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Connected all trees +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO NVLS comm 0x1a308dc0 headRank 3 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Connected all trees +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO NVLS comm 0x19cf06f0 headRank 0 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Connected all trees +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO NVLS comm 0x19250500 headRank 2 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Connected all trees +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO NVLS comm 0x19db9170 headRank 1 nHeads 4 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 805306368 +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO Connected NVLS tree +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-169-239:2207363:2208099 [6] NCCL INFO comm 0x19250500 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0xe68e38d0768d78fc - Init COMPLETE +ip-26-0-169-239:2207362:2208100 [5] NCCL INFO comm 0x19db9170 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe68e38d0768d78fc - Init COMPLETE +ip-26-0-169-239:2207361:2208097 [4] NCCL INFO comm 0x19cf06f0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0xe68e38d0768d78fc - Init COMPLETE +ip-26-0-169-239:2207364:2208098 [7] NCCL INFO comm 0x1a308dc0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0xe68e38d0768d78fc - Init COMPLETE +12/28/2024 01:48:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 4097.70MiB. Peak allocated 6113.75MiB. Peak reserved: 25090.00MiB +12/28/2024 01:48:34 [INFO|DP=0|PP=7|TP=0|ip-26-0-169-239]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 64K | tokens_per_sec: 16.4K | tokens_per_sec_per_gpu: 512 | global_batch_size: 256 | lm_loss: 12 | lr: 0.00015 | model_tflops_per_gpu: 4.95 | hardware_tflops_per_gpu: 4.95 | grad_norm: 0.821 | cuda_memory_allocated: 101K | cuda_max_memory_reserved: 15G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.8G | hd_free_memory_tb: 243G +12/28/2024 01:49:10 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 4097.70MiB. Peak allocated 24411.68MiB. Peak reserved: 25090.00MiB +12/28/2024 01:49:11 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 4097.70MiB. Peak allocated 6113.75MiB. Peak reserved: 25090.00MiB +12/28/2024 01:49:11 [INFO|DP=0|PP=7|TP=0|ip-26-0-169-239]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 36.9K | tokens_per_sec: 28.4K | tokens_per_sec_per_gpu: 887 | global_batch_size: 256 | lm_loss: 12 | lr: 0.0003 | model_tflops_per_gpu: 8.57 | hardware_tflops_per_gpu: 8.57 | grad_norm: 0.82 | cuda_memory_allocated: 101K | cuda_max_memory_reserved: 15G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.8G | hd_free_memory_tb: 243G +12/28/2024 01:49:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 4097.70MiB. Peak allocated 24411.68MiB. Peak reserved: 25090.00MiB +12/28/2024 01:49:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/28/2024 01:49:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: | -------- | -------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | ----------------------------------- | -------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/28/2024 01:49:48 [INFO|DP=0|PP=7|TP=0|ip-26-0-169-239]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 36.7K | tokens_per_sec: 28.6K | tokens_per_sec_per_gpu: 893 | global_batch_size: 256 | lm_loss: 12 | lr: 0.000296 | model_tflops_per_gpu: 8.63 | hardware_tflops_per_gpu: 8.63 | grad_norm: 0.811 | cuda_memory_allocated: 101K | cuda_max_memory_reserved: 15G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.8G | hd_free_memory_tb: 243G +12/28/2024 01:49:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: | 13848463 | 1.34G_dp4_tp1_pp8_acc32_mbs2_seq4096_zero1_tpmodeRED_vocab131k | 4 | 4096 | 2 | 32 | 256 | 8.63 | 8.63 | 892.59 | 166.07 | 97.17 | 95.39 | 422.72 | 242.21 | 242.44 | 5.97 | 24.50 | 4 | 8 | 1 | OneForwardOneBackwardPipelineEngine | REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 32 | 4096 | 131072 | True | torch.bfloat16 | 1 | 25 | True | 1.61G | 470M | +12/28/2024 01:49:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final2.csv +12/28/2024 01:49:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Acquired lock for CSV file: benchmark/results/bench_final2.csv +12/28/2024 01:49:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Successfully wrote to CSV file: benchmark/results/bench_final2.csv. Releasing lock... +12/28/2024 01:49:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Throughput logging complete +ip-26-0-161-138:1967693:1968054 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-161-138:1967692:1968056 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-161-138:1967688:1968058 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-161-138:1967687:1968061 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-161-138:1967693:1967885 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-161-138:1967692:1967886 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-161-138:1967688:1967889 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-161-138:1967687:1967891 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-161-138:1967688:1968058 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-161-138:1967688:1967889 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-169-239:2207361:2208103 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-169-239:2207360:2207821 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-169-239:2207361:2207705 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-169-239:2207359:2207709 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-169-239:2207360:2207706 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-169-239:2207358:2207703 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-169-239:2207361:2207539 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-169-239:2207360:2207543 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-169-239:2207359:2207538 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-169-239:2207358:2207536 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-169-239:2207361:2208103 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-169-239:2207359:2207828 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-169-239:2207361:2207705 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-169-239:2207359:2207709 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-169-239:2207361:2207539 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-169-239:2207359:2207538 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685418:1685922 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-163-220:1685416:1685896 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-163-220:1685414:1685828 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685418:1685789 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685421:1685785 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685417:1685784 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685416:1685787 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685419:1685786 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685414:1685790 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685421:1685618 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685418:1685621 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685417:1685620 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685416:1685616 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685419:1685617 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685414:1685622 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-163-220:1685419:1685888 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-163-220:1685419:1685786 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-163-220:1685418:1685789 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-163-220:1685419:1685617 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-163-220:1685418:1685621 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +[2024-12-28 01:50:05,710] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2207357 closing signal SIGTERM +[2024-12-28 01:50:05,710] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2207358 closing signal SIGTERM +[2024-12-28 01:50:05,710] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2207359 closing signal SIGTERM +[2024-12-28 01:50:05,711] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2207360 closing signal SIGTERM +[2024-12-28 01:50:05,711] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2207361 closing signal SIGTERM +[2024-12-28 01:50:05,711] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2207363 closing signal SIGTERM +[2024-12-28 01:50:10,720] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685414 closing signal SIGTERM +[2024-12-28 01:50:10,720] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685415 closing signal SIGTERM +[2024-12-28 01:50:10,720] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685416 closing signal SIGTERM +[2024-12-28 01:50:10,720] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685417 closing signal SIGTERM +[2024-12-28 01:50:10,720] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685418 closing signal SIGTERM +[2024-12-28 01:50:10,720] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685419 closing signal SIGTERM +[2024-12-28 01:50:10,720] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685421 closing signal SIGTERM +[2024-12-28 01:50:10,726] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1967687 closing signal SIGTERM +[2024-12-28 01:50:10,727] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1967690 closing signal SIGTERM +[2024-12-28 01:50:10,727] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1967692 closing signal SIGTERM +[2024-12-28 01:50:11,104] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1967686) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-28_01:50:10 + host : ip-26-0-161-138.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 1967688) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-28_01:50:10 + host : ip-26-0-161-138.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 1967689) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-28_01:50:10 + host : ip-26-0-161-138.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 1967691) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-28_01:50:10 + host : ip-26-0-161-138.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 1967693) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-28_01:50:10 + host : ip-26-0-161-138.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 1967686) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-161-138: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13848463.0 +[2024-12-28 01:50:11,558] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-28 01:50:11,558] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685415 closing signal SIGTERM +[2024-12-28 01:50:11,558] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685416 closing signal SIGTERM +[2024-12-28 01:50:11,558] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685417 closing signal SIGTERM +[2024-12-28 01:50:11,558] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685418 closing signal SIGTERM +[2024-12-28 01:50:11,558] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685419 closing signal SIGTERM +[2024-12-28 01:50:11,558] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1685421 closing signal SIGTERM +[2024-12-28 01:50:11,558] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-28 01:50:11,558] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-28 01:50:11,559] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2207360 closing signal SIGTERM +[2024-12-28 01:50:11,559] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 940530 closing signal SIGTERM +[2024-12-28 01:50:11,559] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 940531 closing signal SIGTERM +[2024-12-28 01:50:11,559] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 940532 closing signal SIGTERM +[2024-12-28 01:50:11,559] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 940533 closing signal SIGTERM +[2024-12-28 01:50:11,559] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 940534 closing signal SIGTERM +[2024-12-28 01:50:11,559] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 940535 closing signal SIGTERM +[2024-12-28 01:50:11,559] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 940536 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2207286 got signal: 15 +srun: error: ip-26-0-169-239: task 3: Exited with exit code 1 +[2024-12-28 01:50:14,757] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-163-220.ec2.internal_1685341_0' has failed to send a keep-alive heartbeat to the rendezvous '13848463' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 940457 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1685341 got signal: 15 +srun: error: ip-26-0-163-58: task 1: Exited with exit code 1 +srun: error: ip-26-0-163-220: task 2: Exited with exit code 1 diff --git a/logs/13848535-bench_469G_dp4_tp4_pp2_acc64_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13848535-bench_469G_dp4_tp4_pp2_acc64_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..1c45b9a8071cf74148fa8bd84dde69833e2561b2 --- /dev/null +++ b/logs/13848535-bench_469G_dp4_tp4_pp2_acc64_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,12119 @@ ++ '[' -z 13848535 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-162-180,ip-26-0-167-217,ip-26-0-168-[34,95]' ++ export 'NODELIST=ip-26-0-162-180 +ip-26-0-167-217 +ip-26-0-168-34 +ip-26-0-168-95' ++ NODELIST='ip-26-0-162-180 +ip-26-0-167-217 +ip-26-0-168-34 +ip-26-0-168-95' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-162-180,ip-26-0-167-217,ip-26-0-168-[34,95]' ++ export MASTER_NODE=ip-26-0-162-180 ++ MASTER_NODE=ip-26-0-162-180 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-162-180' +Master node: ip-26-0-162-180 ++ echo 'All nodes: ip-26-0-162-180 +ip-26-0-167-217 +ip-26-0-168-34 +ip-26-0-168-95' +All nodes: ip-26-0-162-180 +ip-26-0-167-217 +ip-26-0-168-34 +ip-26-0-168-95 ++ echo 'World size: 32' +World size: 32 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13848535 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-162-180:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_469G_dp4_tp4_pp2_acc64_mbs1_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-28 02:05:05,997] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 02:05:05,997] torch.distributed.run: [WARNING] +[2024-12-28 02:05:05,997] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:05:05,997] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 02:05:05,997] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:05:06,059] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 02:05:06,059] torch.distributed.run: [WARNING] +[2024-12-28 02:05:06,059] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:05:06,059] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 02:05:06,059] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:05:06,136] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 02:05:06,136] torch.distributed.run: [WARNING] +[2024-12-28 02:05:06,136] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:05:06,136] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 02:05:06,136] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:05:06,191] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 02:05:06,191] torch.distributed.run: [WARNING] +[2024-12-28 02:05:06,191] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:05:06,191] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 02:05:06,191] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-162-180:25205:25205 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-162-180:25205:25205 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.162.180<0> +ip-26-0-162-180:25205:25205 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-180:25205:25205 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-180:25205:25205 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-162-180:25212:25212 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-180:25209:25209 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-180:25212:25212 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-162-180:25209:25209 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-162-180:25212:25212 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.162.180<0> +ip-26-0-162-180:25209:25209 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.162.180<0> +ip-26-0-162-180:25212:25212 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-180:25212:25212 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-180:25209:25209 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-180:25209:25209 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-217:138058:138058 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-217:138063:138063 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-217:138063:138063 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-167-217:138058:138058 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-95:1736130:1736130 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1736129:1736129 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1736127:1736127 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1736126:1736126 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1736125:1736125 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-217:138060:138060 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-95:1736128:1736128 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-34:70617:70617 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-34:70613:70613 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-217:138060:138060 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-95:1736128:1736128 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-95:1736130:1736130 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-95:1736125:1736125 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-95:1736126:1736126 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-95:1736127:1736127 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-95:1736129:1736129 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-95:1736124:1736124 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-34:70617:70617 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-95:1736123:1736123 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-34:70613:70613 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-167-217:138058:138058 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.167.217<0> +ip-26-0-167-217:138063:138063 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.167.217<0> +ip-26-0-168-95:1736124:1736124 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-95:1736123:1736123 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-167-217:138060:138060 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.167.217<0> +ip-26-0-167-217:138063:138063 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-217:138058:138058 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-217:138063:138063 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-217:138058:138058 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-180:25206:25206 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-180:25211:25211 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-34:70616:70616 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-180:25206:25206 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-34:70618:70618 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-180:25211:25211 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-34:70616:70616 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-167-217:138060:138060 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-217:138060:138060 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-34:70618:70618 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-162-180:25206:25206 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.162.180<0> +ip-26-0-162-180:25211:25211 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.162.180<0> +ip-26-0-168-34:70615:70615 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-34:70614:70614 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-34:70612:70612 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-34:70617:70617 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.168.34<0> +ip-26-0-162-180:25207:25207 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-168-34:70613:70613 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.168.34<0> +ip-26-0-162-180:25206:25206 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-180:25206:25206 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-34:70611:70611 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-180:25211:25211 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-180:25211:25211 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-34:70615:70615 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-34:70614:70614 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-162-180:25207:25207 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-34:70612:70612 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-168-34:70611:70611 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-162-180:25207:25207 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.162.180<0> +ip-26-0-168-34:70617:70617 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-34:70617:70617 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-34:70613:70613 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-34:70613:70613 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-34:70616:70616 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.168.34<0> +ip-26-0-162-180:25207:25207 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-180:25207:25207 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-34:70618:70618 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.168.34<0> +ip-26-0-168-34:70616:70616 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-34:70616:70616 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1736128:1736128 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1736130:1736130 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1736125:1736125 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1736126:1736126 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1736127:1736127 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1736129:1736129 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1736124:1736124 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-95:1736123:1736123 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.168.95<0> +ip-26-0-168-34:70618:70618 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-34:70618:70618 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-34:70615:70615 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.168.34<0> +ip-26-0-168-34:70614:70614 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.168.34<0> +ip-26-0-168-34:70612:70612 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.168.34<0> +ip-26-0-168-34:70611:70611 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.168.34<0> +ip-26-0-168-95:1736129:1736129 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1736125:1736125 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1736128:1736128 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1736129:1736129 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1736123:1736123 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1736126:1736126 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1736125:1736125 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1736128:1736128 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-34:70615:70615 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-34:70614:70614 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-34:70615:70615 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-34:70614:70614 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1736123:1736123 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1736126:1736126 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1736127:1736127 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1736127:1736127 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-34:70612:70612 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-34:70611:70611 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-34:70612:70612 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-34:70611:70611 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1736130:1736130 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1736124:1736124 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-168-95:1736130:1736130 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-168-95:1736124:1736124 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-217:138056:138056 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-217:138061:138061 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-180:25210:25210 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-217:138056:138056 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-167-217:138062:138062 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-217:138061:138061 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-167-217:138057:138057 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-180:25210:25210 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-167-217:138059:138059 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-167-217:138062:138062 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-167-217:138057:138057 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-162-180:25210:25210 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.162.180<0> +ip-26-0-167-217:138059:138059 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-162-180:25210:25210 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-180:25210:25210 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-217:138056:138056 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.167.217<0> +ip-26-0-167-217:138061:138061 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.167.217<0> +ip-26-0-167-217:138057:138057 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.167.217<0> +ip-26-0-167-217:138062:138062 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.167.217<0> +ip-26-0-167-217:138059:138059 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.167.217<0> +ip-26-0-167-217:138056:138056 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-217:138061:138061 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-217:138061:138061 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-217:138056:138056 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-217:138059:138059 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-217:138059:138059 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-217:138057:138057 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-217:138057:138057 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-167-217:138062:138062 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-167-217:138062:138062 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-180:25208:25208 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-162-180:25208:25208 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-162-180:25208:25208 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.162.180<0> +ip-26-0-162-180:25208:25208 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-162-180:25208:25208 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Using network Libfabric +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-34:70617:70772 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-34:70617:70772 [6] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-180:25209:25366 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-180:25209:25366 [4] NCCL INFO Using network Libfabric +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/217 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/253 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/217 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-217:138058:138217 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-217:138058:138217 [2] NCCL INFO Using network Libfabric +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-217:138057:138221 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-217:138057:138221 [1] NCCL INFO Using network Libfabric +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-217:138060:138219 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-217:138060:138219 [4] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Using network Libfabric +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-34:70612:70778 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-34:70612:70778 [1] NCCL INFO Using network Libfabric +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-34:70615:70775 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-34:70615:70775 [4] NCCL INFO Using network Libfabric +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-34:70611:70773 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-34:70611:70773 [0] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Using network Libfabric +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-34:70613:70774 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-34:70613:70774 [2] NCCL INFO Using network Libfabric +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-34:70614:70777 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-34:70614:70777 [3] NCCL INFO Using network Libfabric +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-34:70618:70771 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-34:70618:70771 [7] NCCL INFO Using network Libfabric +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-168-34:70616:70776 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-168-34:70616:70776 [5] NCCL INFO Using network Libfabric +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-180:25205:25367 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-180:25205:25367 [0] NCCL INFO Using network Libfabric +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-217:138062:138220 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-217:138062:138220 [6] NCCL INFO Using network Libfabric +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-180:25207:25371 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-180:25207:25371 [2] NCCL INFO Using network Libfabric +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-180:25208:25373 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-180:25208:25373 [3] NCCL INFO Using network Libfabric +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-180:25210:25372 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-180:25210:25372 [5] NCCL INFO Using network Libfabric +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-180:25212:25368 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-180:25212:25368 [7] NCCL INFO Using network Libfabric +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-180:25206:25369 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-180:25206:25369 [1] NCCL INFO Using network Libfabric +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-162-180:25211:25370 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-162-180:25211:25370 [6] NCCL INFO Using network Libfabric +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-217:138056:138223 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-217:138056:138223 [0] NCCL INFO Using network Libfabric +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-217:138061:138224 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-217:138061:138224 [5] NCCL INFO Using network Libfabric +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-217:138063:138218 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-217:138063:138218 [7] NCCL INFO Using network Libfabric +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-167-217:138059:138222 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-167-217:138059:138222 [3] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO comm 0x9845e20 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO comm 0x85f36e0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO comm 0x9e68430 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO comm 0x8caf050 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO comm 0xa110da0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO comm 0x90b9930 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-162-180:25210:25372 [5] NCCL INFO comm 0x956d520 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-162-180:25211:25370 [6] NCCL INFO comm 0x9f298b0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-162-180:25212:25368 [7] NCCL INFO comm 0x9245810 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-162-180:25207:25371 [2] NCCL INFO comm 0x87ae030 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-162-180:25206:25369 [1] NCCL INFO comm 0x9ba71e0 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-162-180:25205:25367 [0] NCCL INFO comm 0x94624e0 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-162-180:25208:25373 [3] NCCL INFO comm 0x95a8f60 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-162-180:25209:25366 [4] NCCL INFO comm 0x94e8e50 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-167-217:138063:138218 [7] NCCL INFO comm 0x85f3070 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-167-217:138062:138220 [6] NCCL INFO comm 0x9079b30 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-167-217:138060:138219 [4] NCCL INFO comm 0x89887c0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-167-217:138061:138224 [5] NCCL INFO comm 0x9357b00 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-167-217:138059:138222 [3] NCCL INFO comm 0x9896ea0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO comm 0x9820fc0 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO comm 0xa2e5010 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-34:70618:70771 [7] NCCL INFO comm 0x9664d60 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-34:70615:70775 [4] NCCL INFO comm 0x92a9590 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-34:70611:70773 [0] NCCL INFO comm 0x95efa00 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-167-217:138058:138217 [2] NCCL INFO comm 0x8ece1b0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-167-217:138056:138223 [0] NCCL INFO comm 0x9a6d610 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-167-217:138057:138221 [1] NCCL INFO comm 0xa27a150 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-34:70614:70777 [3] NCCL INFO comm 0x9ce21f0 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-34:70613:70774 [2] NCCL INFO comm 0x968b510 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-34:70617:70772 [6] NCCL INFO comm 0x8956ae0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-34:70616:70776 [5] NCCL INFO comm 0xa409e00 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-168-34:70612:70778 [1] NCCL INFO comm 0x89377a0 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf265befa58b1ac4c - Init START +ip-26-0-167-217:138063:138218 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/217 +ip-26-0-168-34:70612:70778 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70611:70773 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70613:70774 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70614:70777 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138057:138221 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/217 +ip-26-0-167-217:138056:138223 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138058:138217 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-217:138061:138224 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138058:138217 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-168-34:70613:70774 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-34:70613:70774 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-168-34:70617:70772 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70617:70772 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-168-34:70612:70778 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-34:70612:70778 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-167-217:138059:138222 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-217:138059:138222 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-168-34:70618:70771 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70618:70771 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-162-180:25205:25367 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-162-180:25211:25370 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-168-34:70611:70773 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-34:70611:70773 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-162-180:25212:25368 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-162-180:25210:25372 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-167-217:138057:138221 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-217:138057:138221 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-167-217:138060:138219 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138060:138219 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-168-34:70614:70777 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-34:70614:70777 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-168-34:70616:70776 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70616:70776 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-168-34:70615:70775 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70615:70775 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-162-180:25207:25371 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-180:25207:25371 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-162-180:25206:25369 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-162-180:25206:25369 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-167-217:138063:138218 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138063:138218 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-162-180:25209:25366 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-162-180:25208:25373 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-180:25208:25373 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-167-217:138062:138220 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138062:138220 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-167-217:138056:138223 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-217:138056:138223 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Trees [0] 29/-1/-1->28->27 [1] 29/-1/-1->28->27 [2] 29/-1/-1->28->27 [3] 29/-1/-1->28->27 [4] 29/-1/-1->28->20 [5] -1/-1/-1->28->27 [6] 29/-1/-1->28->27 [7] 29/-1/-1->28->27 [8] 29/-1/-1->28->27 [9] 29/-1/-1->28->27 [10] 29/-1/-1->28->27 [11] 29/-1/-1->28->27 [12] 29/12/-1->28->-1 [13] -1/-1/-1->28->27 [14] 29/-1/-1->28->27 [15] 29/-1/-1->28->27 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Trees [0] 30/-1/-1->29->28 [1] 30/-1/-1->29->28 [2] 30/-1/-1->29->28 [3] 30/-1/-1->29->28 [4] 30/-1/-1->29->28 [5] 30/-1/-1->29->21 [6] -1/-1/-1->29->28 [7] 30/-1/-1->29->28 [8] 30/-1/-1->29->28 [9] 30/-1/-1->29->28 [10] 30/-1/-1->29->28 [11] 30/-1/-1->29->28 [12] 30/-1/-1->29->28 [13] 30/13/-1->29->-1 [14] -1/-1/-1->29->28 [15] 30/-1/-1->29->28 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Trees [0] 31/-1/-1->30->29 [1] 31/-1/-1->30->29 [2] 31/-1/-1->30->29 [3] 31/-1/-1->30->29 [4] 31/-1/-1->30->29 [5] 31/-1/-1->30->29 [6] 31/-1/-1->30->22 [7] -1/-1/-1->30->29 [8] 31/-1/-1->30->29 [9] 31/-1/-1->30->29 [10] 31/-1/-1->30->29 [11] 31/-1/-1->30->29 [12] 31/-1/-1->30->29 [13] 31/-1/-1->30->29 [14] 31/14/-1->30->-1 [15] -1/-1/-1->30->29 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25205:25367 [0] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] 24/-1/-1->31->30 [2] 24/-1/-1->31->30 [3] 24/-1/-1->31->30 [4] 24/-1/-1->31->30 [5] 24/-1/-1->31->30 [6] 24/-1/-1->31->30 [7] 24/-1/-1->31->23 [8] -1/-1/-1->31->30 [9] 24/-1/-1->31->30 [10] 24/-1/-1->31->30 [11] 24/-1/-1->31->30 [12] 24/-1/-1->31->30 [13] 24/-1/-1->31->30 [14] 24/-1/-1->31->30 [15] 24/15/-1->31->-1 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25206:25369 [1] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 01/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 03/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 04/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-167-217:138060:138219 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->20 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/20/4->12->28 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 +ip-26-0-167-217:138062:138220 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->22 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/22/6->14->30 [15] -1/-1/-1->14->13 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Trees [0] 26/-1/-1->25->24 [1] 26/-1/-1->25->17 [2] -1/-1/-1->25->24 [3] 26/-1/-1->25->24 [4] 26/-1/-1->25->24 [5] 26/-1/-1->25->24 [6] 26/-1/-1->25->24 [7] 26/-1/-1->25->24 [8] 26/-1/-1->25->24 [9] 26/9/-1->25->-1 [10] -1/-1/-1->25->24 [11] 26/-1/-1->25->24 [12] 26/-1/-1->25->24 [13] 26/-1/-1->25->24 [14] 26/-1/-1->25->24 [15] 26/-1/-1->25->24 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25206:25369 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/17/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 05/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25206:25369 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 06/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-168-34:70613:70774 [2] NCCL INFO Trees [0] 19/-1/-1->18->17 [1] 19/-1/-1->18->17 [2] 19/10/26->18->2 [3] -1/-1/-1->18->17 [4] 19/-1/-1->18->17 [5] 19/-1/-1->18->17 [6] 19/-1/-1->18->17 [7] 19/-1/-1->18->17 [8] 19/-1/-1->18->17 [9] 19/-1/-1->18->17 [10] 19/-1/-1->18->10 [11] -1/-1/-1->18->17 [12] 19/-1/-1->18->17 [13] 19/-1/-1->18->17 [14] 19/-1/-1->18->17 [15] 19/-1/-1->18->17 +ip-26-0-168-34:70611:70773 [0] NCCL INFO Trees [0] 17/8/24->16->0 [1] -1/-1/-1->16->23 [2] 17/-1/-1->16->23 [3] 17/-1/-1->16->23 [4] 17/-1/-1->16->23 [5] 17/-1/-1->16->23 [6] 17/-1/-1->16->23 [7] 17/-1/-1->16->23 [8] 17/-1/-1->16->8 [9] -1/-1/-1->16->23 [10] 17/-1/-1->16->23 [11] 17/-1/-1->16->23 [12] 17/-1/-1->16->23 [13] 17/-1/-1->16->23 [14] 17/-1/-1->16->23 [15] 17/-1/-1->16->23 +ip-26-0-167-217:138056:138223 [0] NCCL INFO Trees [0] 9/-1/-1->8->16 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/16/0->8->24 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 +ip-26-0-167-217:138060:138219 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138062:138220 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138057:138221 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->17 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/17/1->9->25 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 +ip-26-0-167-217:138056:138223 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Trees [0] 27/-1/-1->26->25 [1] 27/-1/-1->26->25 [2] 27/-1/-1->26->18 [3] -1/-1/-1->26->25 [4] 27/-1/-1->26->25 [5] 27/-1/-1->26->25 [6] 27/-1/-1->26->25 [7] 27/-1/-1->26->25 [8] 27/-1/-1->26->25 [9] 27/-1/-1->26->25 [10] 27/10/-1->26->-1 [11] -1/-1/-1->26->25 [12] 27/-1/-1->26->25 [13] 27/-1/-1->26->25 [14] 27/-1/-1->26->25 [15] 27/-1/-1->26->25 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Trees [0] 25/-1/-1->24->16 [1] -1/-1/-1->24->31 [2] 25/-1/-1->24->31 [3] 25/-1/-1->24->31 [4] 25/-1/-1->24->31 [5] 25/-1/-1->24->31 [6] 25/-1/-1->24->31 [7] 25/-1/-1->24->31 [8] 25/8/-1->24->-1 [9] -1/-1/-1->24->31 [10] 25/-1/-1->24->31 [11] 25/-1/-1->24->31 [12] 25/-1/-1->24->31 [13] 25/-1/-1->24->31 [14] 25/-1/-1->24->31 [15] 25/-1/-1->24->31 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 07/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-168-34:70614:70777 [3] NCCL INFO Trees [0] 20/-1/-1->19->18 [1] 20/-1/-1->19->18 [2] 20/-1/-1->19->18 [3] 20/11/27->19->3 [4] -1/-1/-1->19->18 [5] 20/-1/-1->19->18 [6] 20/-1/-1->19->18 [7] 20/-1/-1->19->18 [8] 20/-1/-1->19->18 [9] 20/-1/-1->19->18 [10] 20/-1/-1->19->18 [11] 20/-1/-1->19->11 [12] -1/-1/-1->19->18 [13] 20/-1/-1->19->18 [14] 20/-1/-1->19->18 [15] 20/-1/-1->19->18 +ip-26-0-168-34:70611:70773 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-34:70613:70774 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-34:70614:70777 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138057:138221 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138063:138218 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->23 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/23/7->15->31 +ip-26-0-167-217:138058:138217 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->18 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/18/2->10->26 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 +ip-26-0-167-217:138063:138218 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25208:25373 [3] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 09/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-168-34:70617:70772 [6] NCCL INFO Trees [0] 23/-1/-1->22->21 [1] 23/-1/-1->22->21 [2] 23/-1/-1->22->21 [3] 23/-1/-1->22->21 [4] 23/-1/-1->22->21 [5] 23/-1/-1->22->21 [6] 23/14/30->22->6 [7] -1/-1/-1->22->21 [8] 23/-1/-1->22->21 [9] 23/-1/-1->22->21 [10] 23/-1/-1->22->21 [11] 23/-1/-1->22->21 [12] 23/-1/-1->22->21 [13] 23/-1/-1->22->21 [14] 23/-1/-1->22->14 [15] -1/-1/-1->22->21 +ip-26-0-168-34:70615:70775 [4] NCCL INFO Trees [0] 21/-1/-1->20->19 [1] 21/-1/-1->20->19 [2] 21/-1/-1->20->19 [3] 21/-1/-1->20->19 [4] 21/12/28->20->4 [5] -1/-1/-1->20->19 [6] 21/-1/-1->20->19 [7] 21/-1/-1->20->19 [8] 21/-1/-1->20->19 [9] 21/-1/-1->20->19 [10] 21/-1/-1->20->19 [11] 21/-1/-1->20->19 [12] 21/-1/-1->20->12 [13] -1/-1/-1->20->19 [14] 21/-1/-1->20->19 [15] 21/-1/-1->20->19 +ip-26-0-168-34:70617:70772 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138058:138217 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138061:138224 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->21 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/21/5->13->29 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 +ip-26-0-167-217:138061:138224 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-168-34:70618:70771 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] 16/-1/-1->23->22 [2] 16/-1/-1->23->22 [3] 16/-1/-1->23->22 [4] 16/-1/-1->23->22 [5] 16/-1/-1->23->22 [6] 16/-1/-1->23->22 [7] 16/15/31->23->7 [8] -1/-1/-1->23->22 [9] 16/-1/-1->23->22 [10] 16/-1/-1->23->22 [11] 16/-1/-1->23->22 [12] 16/-1/-1->23->22 [13] 16/-1/-1->23->22 [14] 16/-1/-1->23->22 [15] 16/-1/-1->23->15 +ip-26-0-168-34:70615:70775 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-34:70612:70778 [1] NCCL INFO Trees [0] 18/-1/-1->17->16 [1] 18/9/25->17->1 [2] -1/-1/-1->17->16 [3] 18/-1/-1->17->16 [4] 18/-1/-1->17->16 [5] 18/-1/-1->17->16 [6] 18/-1/-1->17->16 [7] 18/-1/-1->17->16 [8] 18/-1/-1->17->16 [9] 18/-1/-1->17->9 [10] -1/-1/-1->17->16 [11] 18/-1/-1->17->16 [12] 18/-1/-1->17->16 [13] 18/-1/-1->17->16 [14] 18/-1/-1->17->16 [15] 18/-1/-1->17->16 +ip-26-0-167-217:138059:138222 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->19 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/19/3->11->27 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 +ip-26-0-167-217:138059:138222 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25207:25371 [2] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 11/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-168-34:70616:70776 [5] NCCL INFO Trees [0] 22/-1/-1->21->20 [1] 22/-1/-1->21->20 [2] 22/-1/-1->21->20 [3] 22/-1/-1->21->20 [4] 22/-1/-1->21->20 [5] 22/13/29->21->5 [6] -1/-1/-1->21->20 [7] 22/-1/-1->21->20 [8] 22/-1/-1->21->20 [9] 22/-1/-1->21->20 [10] 22/-1/-1->21->20 [11] 22/-1/-1->21->20 [12] 22/-1/-1->21->20 [13] 22/-1/-1->21->13 [14] -1/-1/-1->21->20 [15] 22/-1/-1->21->20 +ip-26-0-168-34:70618:70771 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-34:70612:70778 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-34:70616:70776 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Trees [0] 28/-1/-1->27->26 [1] 28/-1/-1->27->26 [2] 28/-1/-1->27->26 [3] 28/-1/-1->27->19 [4] -1/-1/-1->27->26 [5] 28/-1/-1->27->26 [6] 28/-1/-1->27->26 [7] 28/-1/-1->27->26 [8] 28/-1/-1->27->26 [9] 28/-1/-1->27->26 [10] 28/-1/-1->27->26 [11] 28/11/-1->27->-1 [12] -1/-1/-1->27->26 [13] 28/-1/-1->27->26 [14] 28/-1/-1->27->26 [15] 28/-1/-1->27->26 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25208:25373 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/19/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 +ip-26-0-162-180:25209:25366 [4] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 12/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25210:25372 [5] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25208:25373 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 13/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 14/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-162-180:25211:25370 [6] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 15/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-162-180:25212:25368 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/23/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 +ip-26-0-162-180:25207:25371 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/18/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 +ip-26-0-162-180:25212:25368 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25209:25366 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/20/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 +ip-26-0-162-180:25207:25371 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25210:25372 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/21/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Trees [0] 1/16/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 +ip-26-0-162-180:25209:25366 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25210:25372 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25205:25367 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25211:25370 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/22/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 +ip-26-0-162-180:25211:25370 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 08/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25205:25367 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 08/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 08/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 08/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 08/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 08/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 08/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 08/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 08/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 08/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 08/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 05/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 08/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 06/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 13/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 14/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 05/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 06/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 13/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 14/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 05/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 06/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 13/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 14/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 05/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 06/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 13/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 14/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 08/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 08/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 05/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 05/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 13/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 05/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 13/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 13/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 06/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 06/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 14/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 06/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 07/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 14/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 15/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 07/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 15/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 01/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 14/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 02/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 07/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 03/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 15/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 04/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 09/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 06/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 10/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 14/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 11/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 12/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 07/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 15/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 01/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 02/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 03/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 04/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 09/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 10/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 11/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 12/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 04/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 01/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 01/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 05/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 06/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 04/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 07/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 05/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 04/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 10/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 05/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 11/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 06/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 06/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 12/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 05/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 04/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 07/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 07/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 13/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 05/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 06/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 09/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 09/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 14/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 10/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 09/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 13/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 06/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 06/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 05/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 15/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 11/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 14/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 10/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 07/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 12/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 14/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 13/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 11/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 06/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 06/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 05/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 10/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 14/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 05/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 13/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 12/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 10/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 15/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 11/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 14/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 14/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 13/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 13/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 13/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 12/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 14/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 13/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 15/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 14/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 15/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 07/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 07/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 07/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 15/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 15/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 15/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 02/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 03/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 04/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 05/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 09/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 10/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 11/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 02/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 01/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 12/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 02/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 13/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 04/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 09/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 10/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 12/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 03/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 04/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 05/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 01/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 06/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 03/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 04/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 04/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 07/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 07/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 05/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 09/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 09/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 06/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 09/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 11/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 10/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 07/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 12/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 11/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 10/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 09/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 13/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 12/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 10/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 12/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 14/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 15/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 11/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 11/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 01/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 15/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 14/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 12/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 02/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 13/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 04/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 09/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 10/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 11/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 12/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 01/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 03/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 04/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 05/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 06/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 06/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 07/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 07/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 09/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 09/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 15/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 10/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 11/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 11/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 12/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 14/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 04/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 13/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 07/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 15/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 14/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 09/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 15/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 10/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 11/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 12/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 15/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Connected all rings +ip-26-0-167-217:138056:138223 [0] NCCL INFO Connected all rings +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Connected all rings +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 04/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 05/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Connected all rings +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 06/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Connected all rings +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 07/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 10/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 11/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 12/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 13/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 14/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Connected all rings +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 15/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Connected all rings +ip-26-0-167-217:138060:138219 [4] NCCL INFO Connected all rings +ip-26-0-162-180:25208:25373 [3] NCCL INFO Connected all rings +ip-26-0-168-34:70618:70771 [7] NCCL INFO Connected all rings +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Connected all rings +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Connected all rings +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Connected all rings +ip-26-0-168-34:70611:70773 [0] NCCL INFO Connected all rings +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 04/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 05/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 06/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 07/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Connected all rings +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 10/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 11/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 12/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 13/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Connected all rings +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 14/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 15/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Connected all rings +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 04/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Connected all rings +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 05/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 06/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Connected all rings +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 07/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 09/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 11/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 12/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 13/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Connected all rings +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 14/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 15/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 05/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 06/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Connected all rings +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 07/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 13/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 14/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 04/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Connected all rings +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 15/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 05/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Connected all rings +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Connected all rings +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 06/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 07/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 09/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Connected all rings +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 10/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 12/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 04/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 13/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 05/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 14/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 06/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 15/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 05/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 07/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 06/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 09/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 10/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 07/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 09/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 11/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 11/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 13/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 14/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 15/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 12/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 13/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 04/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 14/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 05/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 06/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Connected all rings +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 15/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 07/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 05/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 09/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Connected all rings +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Connected all rings +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 06/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 10/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 07/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 12/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 13/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 13/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 14/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 14/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Connected all rings +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 15/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 15/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Connected all rings +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Connected all rings +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 00/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 08/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 04/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 05/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Connected all rings +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 05/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 06/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 04/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 06/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 07/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 06/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 09/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 09/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 07/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 10/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 10/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 09/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 11/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 11/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 10/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 13/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 12/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 11/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 14/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 13/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 00/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 12/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 15/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 14/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 08/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 14/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 15/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 04/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 05/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 04/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 04/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 05/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 06/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 07/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 07/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 04/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 07/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 09/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 09/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 05/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 10/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 09/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 10/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 06/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 11/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 10/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 12/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 11/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 09/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 11/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 13/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 15/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 12/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 10/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 12/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 08/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 09/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 13/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 14/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 11/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 15/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 15/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 12/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 08/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 13/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 09/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 14/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 00/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 00/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 03/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 01/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 01/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 03/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 00/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 05/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 00/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 02/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 02/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 05/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 04/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 06/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 04/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 05/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 03/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 03/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 05/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 06/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 08/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 07/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 08/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 08/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 04/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 08/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 04/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 07/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 12/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 08/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 12/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 13/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 05/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 13/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 08/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 05/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 11/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 02/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 06/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 06/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 13/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 11/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 08/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 14/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 07/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 07/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 13/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 10/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 15/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 09/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 09/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 14/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 10/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 15/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 02/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 10/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 08/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 11/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 10/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 12/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 11/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 13/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 14/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 12/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 13/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 15/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 14/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 15/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 06/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 08/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 14/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 06/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 08/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 14/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-162-180:25207:25371 [2] NCCL INFO Connected all trees +ip-26-0-162-180:25207:25371 [2] NCCL INFO NVLS comm 0x87ae030 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25205:25367 [0] NCCL INFO Connected all trees +ip-26-0-162-180:25205:25367 [0] NCCL INFO NVLS comm 0x94624e0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25208:25373 [3] NCCL INFO Connected all trees +ip-26-0-162-180:25208:25373 [3] NCCL INFO NVLS comm 0x95a8f60 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25206:25369 [1] NCCL INFO Connected all trees +ip-26-0-162-180:25212:25368 [7] NCCL INFO Connected all trees +ip-26-0-162-180:25206:25369 [1] NCCL INFO NVLS comm 0x9ba71e0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25212:25368 [7] NCCL INFO NVLS comm 0x9245810 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25211:25370 [6] NCCL INFO Connected all trees +ip-26-0-162-180:25211:25370 [6] NCCL INFO NVLS comm 0x9f298b0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Connected all trees +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Connected all trees +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NVLS comm 0x9820fc0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NVLS comm 0xa110da0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25209:25366 [4] NCCL INFO Connected all trees +ip-26-0-162-180:25209:25366 [4] NCCL INFO NVLS comm 0x94e8e50 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Connected all trees +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NVLS comm 0x8caf050 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Connected all trees +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Connected all trees +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NVLS comm 0xa2e5010 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NVLS comm 0x90b9930 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25210:25372 [5] NCCL INFO Connected all trees +ip-26-0-162-180:25210:25372 [5] NCCL INFO NVLS comm 0x956d520 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Connected all trees +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NVLS comm 0x9845e20 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70616:70776 [5] NCCL INFO Connected all trees +ip-26-0-168-34:70616:70776 [5] NCCL INFO NVLS comm 0xa409e00 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70613:70774 [2] NCCL INFO Connected all trees +ip-26-0-168-34:70613:70774 [2] NCCL INFO NVLS comm 0x968b510 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Connected all trees +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NVLS comm 0x85f36e0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70611:70773 [0] NCCL INFO Connected all trees +ip-26-0-168-34:70611:70773 [0] NCCL INFO NVLS comm 0x95efa00 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70614:70777 [3] NCCL INFO Connected all trees +ip-26-0-168-34:70614:70777 [3] NCCL INFO NVLS comm 0x9ce21f0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70615:70775 [4] NCCL INFO Connected all trees +ip-26-0-168-34:70615:70775 [4] NCCL INFO NVLS comm 0x92a9590 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Connected all trees +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NVLS comm 0x9e68430 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70618:70771 [7] NCCL INFO Connected all trees +ip-26-0-168-34:70618:70771 [7] NCCL INFO NVLS comm 0x9664d60 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138058:138217 [2] NCCL INFO Connected all trees +ip-26-0-167-217:138058:138217 [2] NCCL INFO NVLS comm 0x8ece1b0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138060:138219 [4] NCCL INFO Connected all trees +ip-26-0-167-217:138060:138219 [4] NCCL INFO NVLS comm 0x89887c0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70617:70772 [6] NCCL INFO Connected all trees +ip-26-0-167-217:138063:138218 [7] NCCL INFO Connected all trees +ip-26-0-168-34:70617:70772 [6] NCCL INFO NVLS comm 0x8956ae0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138059:138222 [3] NCCL INFO Connected all trees +ip-26-0-167-217:138061:138224 [5] NCCL INFO Connected all trees +ip-26-0-167-217:138063:138218 [7] NCCL INFO NVLS comm 0x85f3070 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138059:138222 [3] NCCL INFO NVLS comm 0x9896ea0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138061:138224 [5] NCCL INFO NVLS comm 0x9357b00 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138062:138220 [6] NCCL INFO Connected all trees +ip-26-0-167-217:138062:138220 [6] NCCL INFO NVLS comm 0x9079b30 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138056:138223 [0] NCCL INFO Connected all trees +ip-26-0-167-217:138056:138223 [0] NCCL INFO NVLS comm 0x9a6d610 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70612:70778 [1] NCCL INFO Connected all trees +ip-26-0-168-34:70612:70778 [1] NCCL INFO NVLS comm 0x89377a0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138057:138221 [1] NCCL INFO Connected all trees +ip-26-0-167-217:138057:138221 [1] NCCL INFO NVLS comm 0xa27a150 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25372 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25368 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25209:25366 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25211:25370 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:70776 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138224 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:70775 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:70771 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25205:25367 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138057:138221 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138063:138218 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:70778 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138219 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:70772 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25208:25373 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138220 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:70773 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:70777 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138222 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138223 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25371 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138217 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:70774 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25206:25369 [1] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25206:25369 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25206:25369 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25206:25369 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25212:25368 [7] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25212:25368 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25212:25368 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25212:25368 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25207:25371 [2] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25207:25371 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25207:25371 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25207:25371 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25211:25370 [6] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25211:25370 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25211:25370 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25211:25370 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25209:25366 [4] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25209:25366 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25209:25366 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25209:25366 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25205:25367 [0] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25205:25367 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25205:25367 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25205:25367 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25208:25373 [3] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25208:25373 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25208:25373 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25208:25373 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25210:25372 [5] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25210:25372 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25210:25372 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25210:25372 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25211:25370 [6] NCCL INFO comm 0x9f298b0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-162-180:25207:25371 [2] NCCL INFO comm 0x87ae030 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-162-180:25205:25367 [0] NCCL INFO comm 0x94624e0 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-162-180:25209:25366 [4] NCCL INFO comm 0x94e8e50 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-162-180:25210:25372 [5] NCCL INFO comm 0x956d520 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-162-180:25212:25368 [7] NCCL INFO comm 0x9245810 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-162-180:25208:25373 [3] NCCL INFO comm 0x95a8f60 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-162-180:25206:25369 [1] NCCL INFO comm 0x9ba71e0 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-34:70613:70774 [2] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70613:70774 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70613:70774 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70613:70774 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70615:70775 [4] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70615:70775 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70615:70775 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70615:70775 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70611:70773 [0] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70611:70773 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70611:70773 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70611:70773 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70617:70772 [6] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70617:70772 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70617:70772 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70617:70772 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138058:138217 [2] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138058:138217 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138058:138217 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138058:138217 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138062:138220 [6] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138062:138220 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138062:138220 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138062:138220 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138060:138219 [4] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138060:138219 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138060:138219 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138060:138219 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138056:138223 [0] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138056:138223 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138056:138223 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138056:138223 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138063:138218 [7] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138063:138218 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138063:138218 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138063:138218 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70618:70771 [7] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70618:70771 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70618:70771 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70618:70771 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138057:138221 [1] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138057:138221 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138057:138221 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138057:138221 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70612:70778 [1] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70612:70778 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70612:70778 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70612:70778 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138061:138224 [5] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138061:138224 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138061:138224 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138061:138224 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736125:1736283 [2] NCCL INFO comm 0x9820fc0 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-95:1736127:1736290 [4] NCCL INFO comm 0x85f36e0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-95:1736126:1736284 [3] NCCL INFO comm 0xa110da0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-95:1736130:1736288 [7] NCCL INFO comm 0x8caf050 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-95:1736123:1736287 [0] NCCL INFO comm 0x90b9930 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-95:1736124:1736289 [1] NCCL INFO comm 0x9845e20 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-95:1736129:1736285 [6] NCCL INFO comm 0xa2e5010 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-95:1736128:1736286 [5] NCCL INFO comm 0x9e68430 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-34:70614:70777 [3] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70614:70777 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70614:70777 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70614:70777 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70616:70776 [5] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70616:70776 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70616:70776 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70616:70776 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138059:138222 [3] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138059:138222 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138059:138222 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138059:138222 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138058:138217 [2] NCCL INFO comm 0x8ece1b0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-167-217:138060:138219 [4] NCCL INFO comm 0x89887c0 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-167-217:138062:138220 [6] NCCL INFO comm 0x9079b30 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-167-217:138056:138223 [0] NCCL INFO comm 0x9a6d610 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-167-217:138061:138224 [5] NCCL INFO comm 0x9357b00 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-167-217:138063:138218 [7] NCCL INFO comm 0x85f3070 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-167-217:138059:138222 [3] NCCL INFO comm 0x9896ea0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-167-217:138057:138221 [1] NCCL INFO comm 0xa27a150 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-34:70617:70772 [6] NCCL INFO comm 0x8956ae0 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-34:70615:70775 [4] NCCL INFO comm 0x92a9590 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-34:70613:70774 [2] NCCL INFO comm 0x968b510 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-34:70611:70773 [0] NCCL INFO comm 0x95efa00 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-34:70616:70776 [5] NCCL INFO comm 0xa409e00 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-34:70612:70778 [1] NCCL INFO comm 0x89377a0 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-34:70618:70771 [7] NCCL INFO comm 0x9664d60 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-168-34:70614:70777 [3] NCCL INFO comm 0x9ce21f0 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf265befa58b1ac4c - Init COMPLETE +ip-26-0-162-180:25205:25603 [0] NCCL INFO Using network Libfabric +ip-26-0-168-34:70613:71006 [2] NCCL INFO Using network Libfabric +ip-26-0-162-180:25212:25604 [7] NCCL INFO Using network Libfabric +ip-26-0-162-180:25206:25605 [1] NCCL INFO Using network Libfabric +ip-26-0-162-180:25210:25610 [5] NCCL INFO Using network Libfabric +ip-26-0-162-180:25208:25609 [3] NCCL INFO Using network Libfabric +ip-26-0-167-217:138060:138451 [4] NCCL INFO Using network Libfabric +ip-26-0-162-180:25209:25607 [4] NCCL INFO Using network Libfabric +ip-26-0-168-34:70612:71004 [1] NCCL INFO Using network Libfabric +ip-26-0-168-34:70617:71008 [6] NCCL INFO Using network Libfabric +ip-26-0-167-217:138063:138453 [7] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Using network Libfabric +ip-26-0-168-34:70614:71005 [3] NCCL INFO Using network Libfabric +ip-26-0-167-217:138062:138450 [6] NCCL INFO Using network Libfabric +ip-26-0-168-34:70611:71007 [0] NCCL INFO Using network Libfabric +ip-26-0-167-217:138061:138452 [5] NCCL INFO Using network Libfabric +ip-26-0-167-217:138058:138449 [2] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Using network Libfabric +ip-26-0-168-34:70615:71009 [4] NCCL INFO Using network Libfabric +ip-26-0-167-217:138057:138455 [1] NCCL INFO Using network Libfabric +ip-26-0-168-34:70618:71010 [7] NCCL INFO Using network Libfabric +ip-26-0-167-217:138059:138456 [3] NCCL INFO Using network Libfabric +ip-26-0-162-180:25211:25606 [6] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Using network Libfabric +ip-26-0-168-34:70616:71011 [5] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Using network Libfabric +ip-26-0-167-217:138056:138454 [0] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Using network Libfabric +ip-26-0-162-180:25207:25608 [2] NCCL INFO Using network Libfabric +ip-26-0-162-180:25212:25604 [7] NCCL INFO comm 0x947eac0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-162-180:25211:25606 [6] NCCL INFO comm 0xa164880 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-34:70615:71009 [4] NCCL INFO comm 0x94e4960 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-34:70614:71005 [3] NCCL INFO comm 0x9f1cf00 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-34:70613:71006 [2] NCCL INFO comm 0x98c6af0 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-34:70612:71004 [1] NCCL INFO comm 0x8b74c40 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-167-217:138063:138453 [7] NCCL INFO comm 0x882f360 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-167-217:138062:138450 [6] NCCL INFO comm 0x92b48f0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-167-217:138061:138452 [5] NCCL INFO comm 0x9594880 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-167-217:138060:138451 [4] NCCL INFO comm 0x8bc5540 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-162-180:25210:25610 [5] NCCL INFO comm 0x97a85c0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-162-180:25209:25607 [4] NCCL INFO comm 0x9723b40 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-167-217:138059:138456 [3] NCCL INFO comm 0x9ad4ae0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-167-217:138058:138449 [2] NCCL INFO comm 0x910af80 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-167-217:138057:138455 [1] NCCL INFO comm 0xa4b8020 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-162-180:25208:25609 [3] NCCL INFO comm 0x97e2080 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-167-217:138056:138454 [0] NCCL INFO comm 0x9ca9aa0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-162-180:25207:25608 [2] NCCL INFO comm 0x89e7530 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-34:70611:71007 [0] NCCL INFO comm 0x982cc60 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-162-180:25206:25605 [1] NCCL INFO comm 0x9de11f0 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO comm 0xa51de30 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO comm 0x8ee89b0 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-162-180:25205:25603 [0] NCCL INFO comm 0x96a4690 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-34:70617:71008 [6] NCCL INFO comm 0x8b91b00 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-34:70616:71011 [5] NCCL INFO comm 0xa646840 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-34:70618:71010 [7] NCCL INFO comm 0x989f480 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO comm 0xa0a4540 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO comm 0x92f4ae0 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO comm 0xa34adc0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO comm 0x9a80f40 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO comm 0x9a59d20 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO comm 0x882f0c0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa69b8df4f59f0b20 - Init START +ip-26-0-162-180:25212:25604 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25211:25606 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25210:25610 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138057:138455 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138056:138454 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25208:25609 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25209:25607 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70615:71009 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138059:138456 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138058:138449 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/217 +ip-26-0-167-217:138061:138452 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138063:138453 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/217 +ip-26-0-167-217:138060:138451 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25206:25605 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25205:25603 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-162-180:25207:25608 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138062:138450 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70613:71006 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70612:71004 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70614:71005 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70611:71007 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70618:71010 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70617:71008 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70616:71011 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138060:138451 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138060:138451 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-168-34:70615:71009 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70615:71009 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-168-34:70613:71006 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-34:70613:71006 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-167-217:138059:138456 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-217:138059:138456 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-167-217:138058:138449 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-217:138058:138449 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-167-217:138057:138455 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-217:138057:138455 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-168-34:70612:71004 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-34:70612:71004 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-168-34:70617:71008 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70617:71008 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-162-180:25207:25608 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-180:25207:25608 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-162-180:25210:25610 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-162-180:25210:25610 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-168-34:70614:71005 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-34:70614:71005 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-168-34:70618:71010 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70618:71010 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-34:70611:71007 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-34:70611:71007 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-168-34:70616:71011 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70616:71011 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-162-180:25205:25603 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-162-180:25206:25605 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-162-180:25206:25605 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-167-217:138063:138453 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138063:138453 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-167-217:138056:138454 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-217:138056:138454 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-162-180:25212:25604 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-162-180:25212:25604 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-167-217:138061:138452 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138061:138452 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-162-180:25211:25606 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-180:25211:25606 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-167-217:138062:138450 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138062:138450 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-162-180:25208:25609 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-180:25208:25609 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-162-180:25209:25607 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-162-180:25209:25607 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] 24/-1/-1->31->30 [2] 24/-1/-1->31->30 [3] 24/-1/-1->31->30 [4] 24/-1/-1->31->30 [5] 24/-1/-1->31->30 [6] 24/-1/-1->31->30 [7] 24/-1/-1->31->23 [8] -1/-1/-1->31->30 [9] 24/-1/-1->31->30 [10] 24/-1/-1->31->30 [11] 24/-1/-1->31->30 [12] 24/-1/-1->31->30 [13] 24/-1/-1->31->30 [14] 24/-1/-1->31->30 [15] 24/15/-1->31->-1 +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25205:25603 [0] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25205:25603 [0] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25205:25603 [0] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25205:25603 [0] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-162-180:25205:25603 [0] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25205:25603 [0] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25205:25603 [0] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25205:25603 [0] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 01/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 03/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 04/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 05/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 06/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 07/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 09/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 11/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 12/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 13/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 14/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 15/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-162-180:25205:25603 [0] NCCL INFO Trees [0] 1/16/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 +ip-26-0-162-180:25205:25603 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25206:25605 [1] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25206:25605 [1] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25206:25605 [1] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25206:25605 [1] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-162-180:25206:25605 [1] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25206:25605 [1] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25206:25605 [1] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25206:25605 [1] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25206:25605 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/17/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-162-180:25206:25605 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25207:25608 [2] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25207:25608 [2] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25207:25608 [2] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25207:25608 [2] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-167-217:138056:138454 [0] NCCL INFO Trees [0] 9/-1/-1->8->16 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/16/0->8->24 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 +ip-26-0-167-217:138056:138454 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138063:138453 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->23 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/23/7->15->31 +ip-26-0-162-180:25207:25608 [2] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25207:25608 [2] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25207:25608 [2] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25207:25608 [2] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25212:25604 [7] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25212:25604 [7] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25211:25606 [6] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25212:25604 [7] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25211:25606 [6] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25212:25604 [7] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-168-34:70613:71006 [2] NCCL INFO Trees [0] 19/-1/-1->18->17 [1] 19/-1/-1->18->17 [2] 19/10/26->18->2 [3] -1/-1/-1->18->17 [4] 19/-1/-1->18->17 [5] 19/-1/-1->18->17 [6] 19/-1/-1->18->17 [7] 19/-1/-1->18->17 [8] 19/-1/-1->18->17 [9] 19/-1/-1->18->17 [10] 19/-1/-1->18->10 [11] -1/-1/-1->18->17 [12] 19/-1/-1->18->17 [13] 19/-1/-1->18->17 [14] 19/-1/-1->18->17 [15] 19/-1/-1->18->17 +ip-26-0-168-34:70613:71006 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-34:70612:71004 [1] NCCL INFO Trees [0] 18/-1/-1->17->16 [1] 18/9/25->17->1 [2] -1/-1/-1->17->16 [3] 18/-1/-1->17->16 [4] 18/-1/-1->17->16 [5] 18/-1/-1->17->16 [6] 18/-1/-1->17->16 [7] 18/-1/-1->17->16 [8] 18/-1/-1->17->16 [9] 18/-1/-1->17->9 [10] -1/-1/-1->17->16 [11] 18/-1/-1->17->16 [12] 18/-1/-1->17->16 [13] 18/-1/-1->17->16 [14] 18/-1/-1->17->16 [15] 18/-1/-1->17->16 +ip-26-0-167-217:138061:138452 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->21 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/21/5->13->29 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 +ip-26-0-167-217:138063:138453 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138061:138452 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25211:25606 [6] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25212:25604 [7] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25207:25608 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/18/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 +ip-26-0-162-180:25208:25609 [3] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25211:25606 [6] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-162-180:25212:25604 [7] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25210:25610 [5] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25208:25609 [3] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25212:25604 [7] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25207:25608 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138059:138456 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->19 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/19/3->11->27 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 +ip-26-0-167-217:138058:138449 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->18 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/18/2->10->26 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 +ip-26-0-167-217:138059:138456 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138058:138449 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Trees [0] 25/-1/-1->24->16 [1] -1/-1/-1->24->31 [2] 25/-1/-1->24->31 [3] 25/-1/-1->24->31 [4] 25/-1/-1->24->31 [5] 25/-1/-1->24->31 [6] 25/-1/-1->24->31 [7] 25/-1/-1->24->31 [8] 25/8/-1->24->-1 [9] -1/-1/-1->24->31 [10] 25/-1/-1->24->31 [11] 25/-1/-1->24->31 [12] 25/-1/-1->24->31 [13] 25/-1/-1->24->31 [14] 25/-1/-1->24->31 [15] 25/-1/-1->24->31 +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Trees [0] 31/-1/-1->30->29 [1] 31/-1/-1->30->29 [2] 31/-1/-1->30->29 [3] 31/-1/-1->30->29 [4] 31/-1/-1->30->29 [5] 31/-1/-1->30->29 [6] 31/-1/-1->30->22 [7] -1/-1/-1->30->29 [8] 31/-1/-1->30->29 [9] 31/-1/-1->30->29 [10] 31/-1/-1->30->29 [11] 31/-1/-1->30->29 [12] 31/-1/-1->30->29 [13] 31/-1/-1->30->29 [14] 31/14/-1->30->-1 [15] -1/-1/-1->30->29 +ip-26-0-162-180:25211:25606 [6] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25210:25610 [5] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25212:25604 [7] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25210:25610 [5] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25208:25609 [3] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-168-34:70617:71008 [6] NCCL INFO Trees [0] 23/-1/-1->22->21 [1] 23/-1/-1->22->21 [2] 23/-1/-1->22->21 [3] 23/-1/-1->22->21 [4] 23/-1/-1->22->21 [5] 23/-1/-1->22->21 [6] 23/14/30->22->6 [7] -1/-1/-1->22->21 [8] 23/-1/-1->22->21 [9] 23/-1/-1->22->21 [10] 23/-1/-1->22->21 [11] 23/-1/-1->22->21 [12] 23/-1/-1->22->21 [13] 23/-1/-1->22->21 [14] 23/-1/-1->22->14 [15] -1/-1/-1->22->21 +ip-26-0-168-34:70612:71004 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-34:70611:71007 [0] NCCL INFO Trees [0] 17/8/24->16->0 [1] -1/-1/-1->16->23 [2] 17/-1/-1->16->23 [3] 17/-1/-1->16->23 [4] 17/-1/-1->16->23 [5] 17/-1/-1->16->23 [6] 17/-1/-1->16->23 [7] 17/-1/-1->16->23 [8] 17/-1/-1->16->8 [9] -1/-1/-1->16->23 [10] 17/-1/-1->16->23 [11] 17/-1/-1->16->23 [12] 17/-1/-1->16->23 [13] 17/-1/-1->16->23 [14] 17/-1/-1->16->23 [15] 17/-1/-1->16->23 +ip-26-0-168-34:70617:71008 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-34:70611:71007 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138062:138450 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->22 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/22/6->14->30 [15] -1/-1/-1->14->13 +ip-26-0-167-217:138057:138455 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->17 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/17/1->9->25 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 +ip-26-0-167-217:138062:138450 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138057:138455 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25211:25606 [6] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25210:25610 [5] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-162-180:25208:25609 [3] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-162-180:25211:25606 [6] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25209:25607 [4] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-162-180:25211:25606 [6] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25210:25610 [5] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25208:25609 [3] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25210:25610 [5] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25209:25607 [4] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-162-180:25208:25609 [3] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25210:25610 [5] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25209:25607 [4] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-162-180:25208:25609 [3] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-168-34:70615:71009 [4] NCCL INFO Trees [0] 21/-1/-1->20->19 [1] 21/-1/-1->20->19 [2] 21/-1/-1->20->19 [3] 21/-1/-1->20->19 [4] 21/12/28->20->4 [5] -1/-1/-1->20->19 [6] 21/-1/-1->20->19 [7] 21/-1/-1->20->19 [8] 21/-1/-1->20->19 [9] 21/-1/-1->20->19 [10] 21/-1/-1->20->19 [11] 21/-1/-1->20->19 [12] 21/-1/-1->20->12 [13] -1/-1/-1->20->19 [14] 21/-1/-1->20->19 [15] 21/-1/-1->20->19 +ip-26-0-168-34:70615:71009 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-167-217:138060:138451 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->20 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/20/4->12->28 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 +ip-26-0-167-217:138060:138451 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25212:25604 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/23/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 +ip-26-0-162-180:25210:25610 [5] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-162-180:25209:25607 [4] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-162-180:25208:25609 [3] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-168-34:70616:71011 [5] NCCL INFO Trees [0] 22/-1/-1->21->20 [1] 22/-1/-1->21->20 [2] 22/-1/-1->21->20 [3] 22/-1/-1->21->20 [4] 22/-1/-1->21->20 [5] 22/13/29->21->5 [6] -1/-1/-1->21->20 [7] 22/-1/-1->21->20 [8] 22/-1/-1->21->20 [9] 22/-1/-1->21->20 [10] 22/-1/-1->21->20 [11] 22/-1/-1->21->20 [12] 22/-1/-1->21->20 [13] 22/-1/-1->21->13 [14] -1/-1/-1->21->20 [15] 22/-1/-1->21->20 +ip-26-0-168-34:70618:71010 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] 16/-1/-1->23->22 [2] 16/-1/-1->23->22 [3] 16/-1/-1->23->22 [4] 16/-1/-1->23->22 [5] 16/-1/-1->23->22 [6] 16/-1/-1->23->22 [7] 16/15/31->23->7 [8] -1/-1/-1->23->22 [9] 16/-1/-1->23->22 [10] 16/-1/-1->23->22 [11] 16/-1/-1->23->22 [12] 16/-1/-1->23->22 [13] 16/-1/-1->23->22 [14] 16/-1/-1->23->22 [15] 16/-1/-1->23->15 +ip-26-0-168-34:70616:71011 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-34:70618:71010 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25212:25604 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25209:25607 [4] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-162-180:25211:25606 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/22/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 +ip-26-0-162-180:25209:25607 [4] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-162-180:25211:25606 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25209:25607 [4] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-162-180:25209:25607 [4] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-168-34:70614:71005 [3] NCCL INFO Trees [0] 20/-1/-1->19->18 [1] 20/-1/-1->19->18 [2] 20/-1/-1->19->18 [3] 20/11/27->19->3 [4] -1/-1/-1->19->18 [5] 20/-1/-1->19->18 [6] 20/-1/-1->19->18 [7] 20/-1/-1->19->18 [8] 20/-1/-1->19->18 [9] 20/-1/-1->19->18 [10] 20/-1/-1->19->18 [11] 20/-1/-1->19->11 [12] -1/-1/-1->19->18 [13] 20/-1/-1->19->18 [14] 20/-1/-1->19->18 [15] 20/-1/-1->19->18 +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Trees [0] 30/-1/-1->29->28 [1] 30/-1/-1->29->28 [2] 30/-1/-1->29->28 [3] 30/-1/-1->29->28 [4] 30/-1/-1->29->28 [5] 30/-1/-1->29->21 [6] -1/-1/-1->29->28 [7] 30/-1/-1->29->28 [8] 30/-1/-1->29->28 [9] 30/-1/-1->29->28 [10] 30/-1/-1->29->28 [11] 30/-1/-1->29->28 [12] 30/-1/-1->29->28 [13] 30/13/-1->29->-1 [14] -1/-1/-1->29->28 [15] 30/-1/-1->29->28 +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Trees [0] 26/-1/-1->25->24 [1] 26/-1/-1->25->17 [2] -1/-1/-1->25->24 [3] 26/-1/-1->25->24 [4] 26/-1/-1->25->24 [5] 26/-1/-1->25->24 [6] 26/-1/-1->25->24 [7] 26/-1/-1->25->24 [8] 26/-1/-1->25->24 [9] 26/9/-1->25->-1 [10] -1/-1/-1->25->24 [11] 26/-1/-1->25->24 [12] 26/-1/-1->25->24 [13] 26/-1/-1->25->24 [14] 26/-1/-1->25->24 [15] 26/-1/-1->25->24 +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25210:25610 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/21/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-168-34:70614:71005 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25210:25610 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25208:25609 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/19/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 +ip-26-0-162-180:25208:25609 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-162-180:25209:25607 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/20/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 +ip-26-0-162-180:25209:25607 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Trees [0] 27/-1/-1->26->25 [1] 27/-1/-1->26->25 [2] 27/-1/-1->26->18 [3] -1/-1/-1->26->25 [4] 27/-1/-1->26->25 [5] 27/-1/-1->26->25 [6] 27/-1/-1->26->25 [7] 27/-1/-1->26->25 [8] 27/-1/-1->26->25 [9] 27/-1/-1->26->25 [10] 27/10/-1->26->-1 [11] -1/-1/-1->26->25 [12] 27/-1/-1->26->25 [13] 27/-1/-1->26->25 [14] 27/-1/-1->26->25 [15] 27/-1/-1->26->25 +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Trees [0] 29/-1/-1->28->27 [1] 29/-1/-1->28->27 [2] 29/-1/-1->28->27 [3] 29/-1/-1->28->27 [4] 29/-1/-1->28->20 [5] -1/-1/-1->28->27 [6] 29/-1/-1->28->27 [7] 29/-1/-1->28->27 [8] 29/-1/-1->28->27 [9] 29/-1/-1->28->27 [10] 29/-1/-1->28->27 [11] 29/-1/-1->28->27 [12] 29/12/-1->28->-1 [13] -1/-1/-1->28->27 [14] 29/-1/-1->28->27 [15] 29/-1/-1->28->27 +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Trees [0] 28/-1/-1->27->26 [1] 28/-1/-1->27->26 [2] 28/-1/-1->27->26 [3] 28/-1/-1->27->19 [4] -1/-1/-1->27->26 [5] 28/-1/-1->27->26 [6] 28/-1/-1->27->26 [7] 28/-1/-1->27->26 [8] 28/-1/-1->27->26 [9] 28/-1/-1->27->26 [10] 28/-1/-1->27->26 [11] 28/11/-1->27->-1 [12] -1/-1/-1->27->26 [13] 28/-1/-1->27->26 [14] 28/-1/-1->27->26 [15] 28/-1/-1->27->26 +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 08/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 08/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 08/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 08/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 08/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 08/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 08/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 08/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 08/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 08/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 08/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 08/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 05/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 05/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 05/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 06/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 06/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 06/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 13/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 13/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 13/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 14/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 14/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 14/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 05/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 06/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 13/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 14/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 08/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 05/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 13/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 05/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 08/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 13/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 05/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 13/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 06/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 14/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 07/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 06/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 15/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 06/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 14/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 14/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 07/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 07/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 15/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 15/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 01/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 02/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 03/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 04/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 09/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 06/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 10/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 14/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 11/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 12/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 07/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 15/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 01/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 02/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 03/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 04/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 09/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 10/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 11/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 12/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 06/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 01/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 01/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 14/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 09/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 05/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 04/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 04/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 04/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 05/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 05/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 05/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 06/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 13/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 06/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 06/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 06/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 07/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 06/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 07/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 05/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 05/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 05/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 06/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 04/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 07/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 09/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 14/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 13/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 10/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 14/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 13/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 05/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 13/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 09/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 10/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 14/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 10/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 06/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 10/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 11/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 07/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 11/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 11/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 12/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 10/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 12/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 12/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 13/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 13/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 14/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 11/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 13/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 14/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 15/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 12/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 14/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 15/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 13/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 15/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 14/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 15/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 07/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 07/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 07/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 15/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 15/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 02/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 03/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 04/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 05/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 09/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 10/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 11/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 12/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 01/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 13/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 02/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 04/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 09/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 10/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 03/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 12/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 04/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 05/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 01/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 06/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 12/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 15/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 07/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 09/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 06/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 04/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 11/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 07/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 07/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 12/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 09/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 13/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 09/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 10/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 14/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 10/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 11/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 15/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 11/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 14/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 11/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 12/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 15/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 15/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 02/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 03/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 04/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 05/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 09/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 10/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 11/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 12/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 01/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 13/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 02/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 04/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 09/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 03/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 10/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 04/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 12/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 05/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 01/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 06/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 04/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 07/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 07/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 09/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 06/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 07/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 09/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 11/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 09/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 10/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 12/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 10/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 11/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 13/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 14/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 11/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 14/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 12/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 15/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 15/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 15/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Connected all rings +ip-26-0-167-217:138056:138454 [0] NCCL INFO Connected all rings +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 04/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 05/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 06/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 07/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 09/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 10/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 12/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 13/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 14/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 15/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Connected all rings +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 04/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 05/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 06/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 07/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 09/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 11/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 12/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 13/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 14/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 15/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Connected all rings +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 04/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 05/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 06/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 07/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 10/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 11/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 12/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 13/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 14/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 15/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 05/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 06/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 07/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 13/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 14/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 15/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Connected all rings +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Connected all rings +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Connected all rings +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Connected all rings +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Connected all rings +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Connected all rings +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Connected all rings +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Connected all rings +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Connected all rings +ip-26-0-162-180:25208:25609 [3] NCCL INFO Connected all rings +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 05/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 06/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 07/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 09/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 10/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 11/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 13/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 14/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 15/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Connected all rings +ip-26-0-162-180:25209:25607 [4] NCCL INFO Connected all rings +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Connected all rings +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Connected all rings +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Connected all rings +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 04/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 05/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 07/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 09/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 10/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 11/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 12/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 13/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 15/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Connected all rings +ip-26-0-168-34:70615:71009 [4] NCCL INFO Connected all rings +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 04/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 04/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 06/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 07/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 09/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 10/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 11/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 12/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 14/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 15/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Connected all rings +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Connected all rings +ip-26-0-162-180:25206:25605 [1] NCCL INFO Connected all rings +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 05/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 06/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 07/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 10/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 11/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 12/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Connected all rings +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 13/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Connected all rings +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 14/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 15/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 04/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 05/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 06/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 09/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 10/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 11/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 12/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 13/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 14/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Connected all rings +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 04/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 05/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 06/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 07/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Connected all rings +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 09/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 11/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 12/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 13/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 14/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 15/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 05/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 06/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 07/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 13/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 14/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 15/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Connected all rings +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Connected all rings +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Connected all rings +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 04/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 05/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 06/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 07/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 09/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 10/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 12/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 13/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Connected all rings +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 14/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 05/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 15/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 04/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 05/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 06/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 06/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 04/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 07/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 09/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 06/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 09/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 10/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 07/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 10/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 00/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 11/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 09/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 11/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 12/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 10/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 13/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 00/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 08/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 13/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 11/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 14/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 08/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 14/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 12/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 15/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 14/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 15/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 04/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 05/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 07/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 09/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 10/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 11/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 01/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 01/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 02/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 08/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 02/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 10/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 03/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 02/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 12/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 04/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 13/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 03/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 08/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 05/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 09/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 06/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 15/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 04/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 07/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 05/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 08/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 06/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 09/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 09/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 07/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 09/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 02/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 10/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 08/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 11/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 10/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 10/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 12/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 00/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 00/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 13/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 05/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 03/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 04/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 00/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 11/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 00/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 14/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 03/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 12/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 04/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 08/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 08/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 05/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 15/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 13/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 05/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 05/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 08/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 13/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 06/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 12/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 06/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 08/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 14/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 12/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 07/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 06/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 08/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 07/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 13/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 15/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 08/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 11/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 08/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 14/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 11/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 06/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 13/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 13/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 08/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 14/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 14/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 14/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 15/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 15/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25205:25603 [0] NCCL INFO Connected all trees +ip-26-0-162-180:25205:25603 [0] NCCL INFO NVLS comm 0x96a4690 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Connected all trees +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO NVLS comm 0x9a59d20 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Connected all trees +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO NVLS comm 0xa51de30 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Connected all trees +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO NVLS comm 0x92f4ae0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25209:25607 [4] NCCL INFO Connected all trees +ip-26-0-162-180:25209:25607 [4] NCCL INFO NVLS comm 0x9723b40 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25212:25604 [7] NCCL INFO Connected all trees +ip-26-0-162-180:25212:25604 [7] NCCL INFO NVLS comm 0x947eac0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25207:25608 [2] NCCL INFO Connected all trees +ip-26-0-162-180:25207:25608 [2] NCCL INFO NVLS comm 0x89e7530 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Connected all trees +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO NVLS comm 0xa34adc0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Connected all trees +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO NVLS comm 0x882f0c0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Connected all trees +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Connected all trees +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO NVLS comm 0x8ee89b0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO NVLS comm 0xa0a4540 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25208:25609 [3] NCCL INFO Connected all trees +ip-26-0-162-180:25208:25609 [3] NCCL INFO NVLS comm 0x97e2080 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Connected all trees +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO NVLS comm 0x9a80f40 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25210:25610 [5] NCCL INFO Connected all trees +ip-26-0-162-180:25210:25610 [5] NCCL INFO NVLS comm 0x97a85c0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25206:25605 [1] NCCL INFO Connected all trees +ip-26-0-162-180:25206:25605 [1] NCCL INFO NVLS comm 0x9de11f0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70611:71007 [0] NCCL INFO Connected all trees +ip-26-0-168-34:70611:71007 [0] NCCL INFO NVLS comm 0x982cc60 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70618:71010 [7] NCCL INFO Connected all trees +ip-26-0-168-34:70612:71004 [1] NCCL INFO Connected all trees +ip-26-0-168-34:70618:71010 [7] NCCL INFO NVLS comm 0x989f480 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70612:71004 [1] NCCL INFO NVLS comm 0x8b74c40 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70613:71006 [2] NCCL INFO Connected all trees +ip-26-0-168-34:70613:71006 [2] NCCL INFO NVLS comm 0x98c6af0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25211:25606 [6] NCCL INFO Connected all trees +ip-26-0-162-180:25211:25606 [6] NCCL INFO NVLS comm 0xa164880 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70617:71008 [6] NCCL INFO Connected all trees +ip-26-0-168-34:70617:71008 [6] NCCL INFO NVLS comm 0x8b91b00 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138058:138449 [2] NCCL INFO Connected all trees +ip-26-0-167-217:138058:138449 [2] NCCL INFO NVLS comm 0x910af80 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138063:138453 [7] NCCL INFO Connected all trees +ip-26-0-167-217:138063:138453 [7] NCCL INFO NVLS comm 0x882f360 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70615:71009 [4] NCCL INFO Connected all trees +ip-26-0-168-34:70615:71009 [4] NCCL INFO NVLS comm 0x94e4960 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70614:71005 [3] NCCL INFO Connected all trees +ip-26-0-168-34:70614:71005 [3] NCCL INFO NVLS comm 0x9f1cf00 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138056:138454 [0] NCCL INFO Connected all trees +ip-26-0-167-217:138056:138454 [0] NCCL INFO NVLS comm 0x9ca9aa0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70616:71011 [5] NCCL INFO Connected all trees +ip-26-0-168-34:70616:71011 [5] NCCL INFO NVLS comm 0xa646840 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138057:138455 [1] NCCL INFO Connected all trees +ip-26-0-167-217:138057:138455 [1] NCCL INFO NVLS comm 0xa4b8020 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138059:138456 [3] NCCL INFO Connected all trees +ip-26-0-167-217:138059:138456 [3] NCCL INFO NVLS comm 0x9ad4ae0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138062:138450 [6] NCCL INFO Connected all trees +ip-26-0-167-217:138062:138450 [6] NCCL INFO NVLS comm 0x92b48f0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138061:138452 [5] NCCL INFO Connected all trees +ip-26-0-167-217:138061:138452 [5] NCCL INFO NVLS comm 0x9594880 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138060:138451 [4] NCCL INFO Connected all trees +ip-26-0-167-217:138060:138451 [4] NCCL INFO NVLS comm 0x8bc5540 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25210:25610 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25209:25607 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25212:25604 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138061:138452 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70616:71011 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138059:138456 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138063:138453 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138060:138451 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-167-217:138056:138454 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70615:71009 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70617:71008 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-167-217:138057:138455 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-167-217:138058:138449 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-167-217:138062:138450 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70618:71010 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-168-34:70611:71007 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70613:71006 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70612:71004 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-168-34:70614:71005 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25208:25609 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25206:25605 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-162-180:25205:25603 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25211:25606 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-162-180:25207:25608 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25210:25610 [5] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25210:25610 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25210:25610 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25210:25610 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25212:25604 [7] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25212:25604 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25212:25604 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25212:25604 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25206:25605 [1] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25206:25605 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25206:25605 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25206:25605 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25208:25609 [3] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25208:25609 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25208:25609 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25208:25609 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-95:1736128:1736517 [5] NCCL INFO comm 0xa0a4540 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-95:1736126:1736522 [3] NCCL INFO comm 0xa34adc0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-95:1736124:1736521 [1] NCCL INFO comm 0x9a80f40 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-95:1736130:1736516 [7] NCCL INFO comm 0x8ee89b0 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-95:1736129:1736518 [6] NCCL INFO comm 0xa51de30 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-95:1736125:1736523 [2] NCCL INFO comm 0x9a59d20 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-95:1736123:1736519 [0] NCCL INFO comm 0x92f4ae0 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-95:1736127:1736520 [4] NCCL INFO comm 0x882f0c0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-167-217:138057:138455 [1] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138057:138455 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138057:138455 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138057:138455 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138059:138456 [3] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138059:138456 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138059:138456 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138059:138456 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138063:138453 [7] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138063:138453 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138063:138453 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138063:138453 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25211:25606 [6] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25211:25606 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25211:25606 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25211:25606 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70617:71008 [6] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70617:71008 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70617:71008 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70617:71008 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70612:71004 [1] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70612:71004 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70612:71004 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70612:71004 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70616:71011 [5] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70616:71011 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70616:71011 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70616:71011 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138061:138452 [5] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138061:138452 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138061:138452 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138061:138452 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70618:71010 [7] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70618:71010 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70618:71010 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70618:71010 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138062:138450 [6] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138062:138450 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138062:138450 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138062:138450 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25207:25608 [2] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25207:25608 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25207:25608 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25207:25608 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25209:25607 [4] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25209:25607 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25209:25607 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25209:25607 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70615:71009 [4] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70615:71009 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70615:71009 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70615:71009 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70614:71005 [3] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70614:71005 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70614:71005 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70614:71005 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25205:25603 [0] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25205:25603 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25205:25603 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-162-180:25205:25603 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70611:71007 [0] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70611:71007 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70611:71007 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70611:71007 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-168-34:70613:71006 [2] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70613:71006 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70613:71006 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-168-34:70613:71006 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138058:138449 [2] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138058:138449 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138058:138449 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138058:138449 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138060:138451 [4] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138060:138451 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138060:138451 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138060:138451 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-167-217:138056:138454 [0] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138056:138454 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138056:138454 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-167-217:138056:138454 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-162-180:25210:25610 [5] NCCL INFO comm 0x97a85c0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-162-180:25208:25609 [3] NCCL INFO comm 0x97e2080 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-162-180:25212:25604 [7] NCCL INFO comm 0x947eac0 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-162-180:25209:25607 [4] NCCL INFO comm 0x9723b40 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-162-180:25211:25606 [6] NCCL INFO comm 0xa164880 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-162-180:25207:25608 [2] NCCL INFO comm 0x89e7530 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-162-180:25206:25605 [1] NCCL INFO comm 0x9de11f0 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-162-180:25205:25603 [0] NCCL INFO comm 0x96a4690 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-34:70618:71010 [7] NCCL INFO comm 0x989f480 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-34:70616:71011 [5] NCCL INFO comm 0xa646840 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-34:70612:71004 [1] NCCL INFO comm 0x8b74c40 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-34:70614:71005 [3] NCCL INFO comm 0x9f1cf00 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-34:70615:71009 [4] NCCL INFO comm 0x94e4960 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-34:70617:71008 [6] NCCL INFO comm 0x8b91b00 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-34:70613:71006 [2] NCCL INFO comm 0x98c6af0 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-168-34:70611:71007 [0] NCCL INFO comm 0x982cc60 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-167-217:138058:138449 [2] NCCL INFO comm 0x910af80 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-167-217:138061:138452 [5] NCCL INFO comm 0x9594880 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-167-217:138059:138456 [3] NCCL INFO comm 0x9ad4ae0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-167-217:138056:138454 [0] NCCL INFO comm 0x9ca9aa0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-167-217:138063:138453 [7] NCCL INFO comm 0x882f360 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-167-217:138060:138451 [4] NCCL INFO comm 0x8bc5540 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-167-217:138062:138450 [6] NCCL INFO comm 0x92b48f0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-167-217:138057:138455 [1] NCCL INFO comm 0xa4b8020 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xa69b8df4f59f0b20 - Init COMPLETE +ip-26-0-162-180:25210:25644 [5] NCCL INFO Using network Libfabric +ip-26-0-162-180:25205:25640 [0] NCCL INFO Using network Libfabric +ip-26-0-162-180:25211:25641 [6] NCCL INFO Using network Libfabric +ip-26-0-162-180:25208:25642 [3] NCCL INFO Using network Libfabric +ip-26-0-162-180:25206:25647 [1] NCCL INFO Using network Libfabric +ip-26-0-162-180:25212:25643 [7] NCCL INFO Using network Libfabric +ip-26-0-162-180:25209:25645 [4] NCCL INFO Using network Libfabric +ip-26-0-162-180:25207:25646 [2] NCCL INFO Using network Libfabric +ip-26-0-162-180:25205:25640 [0] NCCL INFO comm 0x96bb550 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5e3dea37485e3e97 - Init START +ip-26-0-162-180:25206:25647 [1] NCCL INFO comm 0x9df86d0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5e3dea37485e3e97 - Init START +ip-26-0-162-180:25212:25643 [7] NCCL INFO comm 0x9494a60 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5e3dea37485e3e97 - Init START +ip-26-0-162-180:25208:25642 [3] NCCL INFO comm 0x97f9da0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5e3dea37485e3e97 - Init START +ip-26-0-162-180:25207:25646 [2] NCCL INFO comm 0x89fe720 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5e3dea37485e3e97 - Init START +ip-26-0-162-180:25211:25641 [6] NCCL INFO comm 0xa17a3b0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5e3dea37485e3e97 - Init START +ip-26-0-162-180:25209:25645 [4] NCCL INFO comm 0x9739ef0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5e3dea37485e3e97 - Init START +ip-26-0-162-180:25210:25644 [5] NCCL INFO comm 0x97bca90 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5e3dea37485e3e97 - Init START +ip-26-0-162-180:25206:25647 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25212:25643 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25205:25640 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-162-180:25207:25646 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25208:25642 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25211:25641 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25209:25645 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-162-180:25210:25644 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Using network Libfabric +ip-26-0-168-34:70616:71048 [5] NCCL INFO Using network Libfabric +ip-26-0-168-34:70613:71046 [2] NCCL INFO Using network Libfabric +ip-26-0-168-34:70615:71047 [4] NCCL INFO Using network Libfabric +ip-26-0-168-34:70612:71045 [1] NCCL INFO Using network Libfabric +ip-26-0-168-34:70618:71044 [7] NCCL INFO Using network Libfabric +ip-26-0-168-34:70614:71049 [3] NCCL INFO Using network Libfabric +ip-26-0-168-34:70617:71043 [6] NCCL INFO Using network Libfabric +ip-26-0-168-34:70611:71042 [0] NCCL INFO comm 0x9843040 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb445fc8e45153ce3 - Init START +ip-26-0-168-34:70612:71045 [1] NCCL INFO comm 0x8b8abf0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb445fc8e45153ce3 - Init START +ip-26-0-168-34:70615:71047 [4] NCCL INFO comm 0x94fbd90 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb445fc8e45153ce3 - Init START +ip-26-0-168-34:70613:71046 [2] NCCL INFO comm 0x98dee90 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb445fc8e45153ce3 - Init START +ip-26-0-168-34:70618:71044 [7] NCCL INFO comm 0x98b7520 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb445fc8e45153ce3 - Init START +ip-26-0-168-34:70614:71049 [3] NCCL INFO comm 0x9f34f60 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb445fc8e45153ce3 - Init START +ip-26-0-168-34:70617:71043 [6] NCCL INFO comm 0x8ba9050 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb445fc8e45153ce3 - Init START +ip-26-0-168-34:70616:71048 [5] NCCL INFO comm 0xa65d6d0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb445fc8e45153ce3 - Init START +ip-26-0-168-34:70615:71047 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70614:71049 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70613:71046 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70612:71045 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70618:71044 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70611:71042 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70616:71048 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-34:70617:71043 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Using network Libfabric +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO comm 0xa0ba010 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x14693696db1c5513 - Init START +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO comm 0x930a490 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x14693696db1c5513 - Init START +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO comm 0x9a71010 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x14693696db1c5513 - Init START +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO comm 0x8844760 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x14693696db1c5513 - Init START +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO comm 0x8eff3f0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x14693696db1c5513 - Init START +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO comm 0x9a96820 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x14693696db1c5513 - Init START +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO comm 0xa361700 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x14693696db1c5513 - Init START +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO comm 0xa5359c0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x14693696db1c5513 - Init START +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Using network Libfabric +ip-26-0-167-217:138059:138491 [3] NCCL INFO Using network Libfabric +ip-26-0-167-217:138062:138490 [6] NCCL INFO Using network Libfabric +ip-26-0-167-217:138063:138492 [7] NCCL INFO Using network Libfabric +ip-26-0-167-217:138060:138488 [4] NCCL INFO Using network Libfabric +ip-26-0-167-217:138058:138494 [2] NCCL INFO Using network Libfabric +ip-26-0-167-217:138057:138493 [1] NCCL INFO Using network Libfabric +ip-26-0-167-217:138061:138489 [5] NCCL INFO Using network Libfabric +ip-26-0-167-217:138056:138487 [0] NCCL INFO comm 0x9cc0900 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb27c5bb1aa7bfe6 - Init START +ip-26-0-167-217:138057:138493 [1] NCCL INFO comm 0xa4cdb00 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb27c5bb1aa7bfe6 - Init START +ip-26-0-167-217:138063:138492 [7] NCCL INFO comm 0x8845650 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb27c5bb1aa7bfe6 - Init START +ip-26-0-167-217:138061:138489 [5] NCCL INFO comm 0x95aa710 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb27c5bb1aa7bfe6 - Init START +ip-26-0-167-217:138062:138490 [6] NCCL INFO comm 0x92cd910 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb27c5bb1aa7bfe6 - Init START +ip-26-0-167-217:138058:138494 [2] NCCL INFO comm 0x9120a80 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb27c5bb1aa7bfe6 - Init START +ip-26-0-167-217:138060:138488 [4] NCCL INFO comm 0x8bdd0a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb27c5bb1aa7bfe6 - Init START +ip-26-0-167-217:138059:138491 [3] NCCL INFO comm 0x9aeb140 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb27c5bb1aa7bfe6 - Init START +ip-26-0-167-217:138056:138487 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138063:138492 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/217 +ip-26-0-167-217:138062:138490 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138057:138493 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138061:138489 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138060:138488 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138058:138494 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/217 +ip-26-0-167-217:138059:138491 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-167-217:138061:138489 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138061:138489 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-168-34:70614:71049 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-34:70614:71049 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-167-217:138059:138491 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-167-217:138059:138491 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-167-217:138058:138494 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-167-217:138058:138494 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-168-34:70611:71042 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-217:138056:138487 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-167-217:138062:138490 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138062:138490 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-167-217:138060:138488 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138060:138488 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-167-217:138063:138492 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-167-217:138063:138492 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-34:70612:71045 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-34:70612:71045 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-168-34:70613:71046 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-34:70613:71046 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-167-217:138057:138493 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-167-217:138057:138493 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-167-217:138057:138493 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-167-217:138057:138493 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-167-217:138058:138494 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-167-217:138058:138494 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-167-217:138060:138488 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-167-217:138061:138489 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138060:138488 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-167-217:138061:138489 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138062:138490 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138062:138490 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138063:138492 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138063:138492 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138059:138491 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138059:138491 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-167-217:138056:138487 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-34:70617:71043 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70617:71043 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-168-34:70615:71047 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70615:71047 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70616:71048 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-34:70618:71044 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-168-34:70618:71044 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-168-34:70618:71044 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70613:71046 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70612:71045 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-168-34:70613:71046 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70612:71045 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70617:71043 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-168-34:70615:71047 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70617:71043 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-34:70616:71048 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-168-34:70615:71047 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70616:71048 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70614:71049 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-168-34:70611:71042 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-34:70614:71049 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-162-180:25212:25643 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-162-180:25206:25647 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-162-180:25210:25644 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-162-180:25207:25646 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-162-180:25205:25640 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-162-180:25208:25642 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-162-180:25208:25642 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-162-180:25211:25641 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-162-180:25207:25646 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-162-180:25208:25642 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-162-180:25209:25645 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-162-180:25210:25644 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-162-180:25207:25646 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-162-180:25208:25642 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-162-180:25210:25644 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-162-180:25209:25645 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-162-180:25206:25647 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-162-180:25206:25647 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25211:25641 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-162-180:25212:25643 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25211:25641 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-162-180:25212:25643 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-162-180:25205:25640 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Connected all rings +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Connected all rings +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Connected all rings +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Connected all rings +ip-26-0-167-217:138062:138490 [6] NCCL INFO Connected all rings +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Connected all rings +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Connected all rings +ip-26-0-167-217:138057:138493 [1] NCCL INFO Connected all rings +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Connected all rings +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Connected all rings +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Connected all rings +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Connected all rings +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Connected all rings +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Connected all rings +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138060:138488 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Connected all rings +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Connected all rings +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138057:138493 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Connected all rings +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70612:71045 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70616:71048 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Connected all rings +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Connected all rings +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70615:71047 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-167-217:138058:138494 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70617:71043 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Connected all rings +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70614:71049 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Connected all rings +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Connected all rings +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-34:70613:71046 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138059:138491 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Connected all rings +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Connected all rings +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Connected all rings +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Connected all rings +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO Connected all trees +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-34:70611:71042 [0] NCCL INFO NVLS comm 0x9843040 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO Connected all trees +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138063:138492 [7] NCCL INFO NVLS comm 0x8845650 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-167-217:138056:138487 [0] NCCL INFO Connected all trees +ip-26-0-167-217:138056:138487 [0] NCCL INFO NVLS comm 0x9cc0900 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Connected all rings +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Connected all rings +ip-26-0-162-180:25206:25647 [1] NCCL INFO Connected all rings +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Connected all rings +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Connected all rings +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25205:25640 [0] NCCL INFO Connected all rings +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Connected all trees +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO NVLS comm 0x930a490 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25212:25643 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25207:25646 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25208:25642 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25210:25644 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25206:25647 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-168-34:70618:71044 [7] NCCL INFO Connected all trees +ip-26-0-168-34:70618:71044 [7] NCCL INFO NVLS comm 0x98b7520 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO Connected all trees +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138061:138489 [5] NCCL INFO NVLS comm 0x95aa710 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25211:25641 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-162-180:25209:25645 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-167-217:138062:138490 [6] NCCL INFO Connected all trees +ip-26-0-167-217:138062:138490 [6] NCCL INFO NVLS comm 0x92cd910 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Connected all trees +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO NVLS comm 0x8eff3f0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25205:25640 [0] NCCL INFO Connected all trees +ip-26-0-162-180:25205:25640 [0] NCCL INFO NVLS comm 0x96bb550 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138057:138493 [1] NCCL INFO Connected all trees +ip-26-0-167-217:138057:138493 [1] NCCL INFO NVLS comm 0xa4cdb00 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138059:138491 [3] NCCL INFO Connected all trees +ip-26-0-167-217:138059:138491 [3] NCCL INFO NVLS comm 0x9aeb140 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138058:138494 [2] NCCL INFO Connected all trees +ip-26-0-167-217:138058:138494 [2] NCCL INFO NVLS comm 0x9120a80 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138060:138488 [4] NCCL INFO Connected all trees +ip-26-0-167-217:138060:138488 [4] NCCL INFO NVLS comm 0x8bdd0a0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70613:71046 [2] NCCL INFO Connected all trees +ip-26-0-168-34:70613:71046 [2] NCCL INFO NVLS comm 0x98dee90 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70616:71048 [5] NCCL INFO Connected all trees +ip-26-0-168-34:70616:71048 [5] NCCL INFO NVLS comm 0xa65d6d0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70615:71047 [4] NCCL INFO Connected all trees +ip-26-0-168-34:70615:71047 [4] NCCL INFO NVLS comm 0x94fbd90 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25212:25643 [7] NCCL INFO Connected all trees +ip-26-0-162-180:25212:25643 [7] NCCL INFO NVLS comm 0x9494a60 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70617:71043 [6] NCCL INFO Connected all trees +ip-26-0-168-34:70617:71043 [6] NCCL INFO NVLS comm 0x8ba9050 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70612:71045 [1] NCCL INFO Connected all trees +ip-26-0-168-34:70612:71045 [1] NCCL INFO NVLS comm 0x8b8abf0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70614:71049 [3] NCCL INFO Connected all trees +ip-26-0-168-34:70614:71049 [3] NCCL INFO NVLS comm 0x9f34f60 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-167-217:138061:138489 [5] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138061:138489 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138061:138489 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-217:138061:138489 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-217:138057:138493 [1] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138057:138493 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138057:138493 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-217:138057:138493 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-217:138059:138491 [3] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138059:138491 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138059:138491 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-217:138059:138491 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-217:138060:138488 [4] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138060:138488 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138060:138488 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-217:138060:138488 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-217:138058:138494 [2] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138056:138487 [0] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138056:138487 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138058:138494 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138058:138494 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-217:138058:138494 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-217:138056:138487 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-217:138056:138487 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-217:138062:138490 [6] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138062:138490 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138062:138490 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-217:138062:138490 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-217:138063:138492 [7] NCCL INFO Connected NVLS tree +ip-26-0-167-217:138063:138492 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-167-217:138063:138492 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-167-217:138063:138492 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-167-217:138062:138490 [6] NCCL INFO comm 0x92cd910 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb27c5bb1aa7bfe6 - Init COMPLETE +ip-26-0-167-217:138058:138494 [2] NCCL INFO comm 0x9120a80 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb27c5bb1aa7bfe6 - Init COMPLETE +ip-26-0-167-217:138056:138487 [0] NCCL INFO comm 0x9cc0900 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb27c5bb1aa7bfe6 - Init COMPLETE +ip-26-0-167-217:138060:138488 [4] NCCL INFO comm 0x8bdd0a0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb27c5bb1aa7bfe6 - Init COMPLETE +ip-26-0-167-217:138061:138489 [5] NCCL INFO comm 0x95aa710 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb27c5bb1aa7bfe6 - Init COMPLETE +ip-26-0-167-217:138059:138491 [3] NCCL INFO comm 0x9aeb140 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb27c5bb1aa7bfe6 - Init COMPLETE +ip-26-0-167-217:138063:138492 [7] NCCL INFO comm 0x8845650 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb27c5bb1aa7bfe6 - Init COMPLETE +ip-26-0-167-217:138057:138493 [1] NCCL INFO comm 0xa4cdb00 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb27c5bb1aa7bfe6 - Init COMPLETE +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Connected all trees +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO NVLS comm 0xa0ba010 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Connected all trees +ip-26-0-168-34:70613:71046 [2] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70613:71046 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70613:71046 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-34:70613:71046 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-34:70617:71043 [6] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70617:71043 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70617:71043 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-34:70617:71043 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO NVLS comm 0xa5359c0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70615:71047 [4] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70615:71047 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70611:71042 [0] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70615:71047 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-34:70615:71047 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-34:70611:71042 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70611:71042 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-34:70611:71042 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-34:70612:71045 [1] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70612:71045 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70612:71045 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-34:70612:71045 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-34:70618:71044 [7] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70618:71044 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70618:71044 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-34:70618:71044 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-34:70616:71048 [5] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70616:71048 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70616:71048 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-34:70616:71048 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-34:70614:71049 [3] NCCL INFO Connected NVLS tree +ip-26-0-168-34:70614:71049 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-34:70614:71049 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-34:70614:71049 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Connected all trees +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO NVLS comm 0x8844760 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-34:70615:71047 [4] NCCL INFO comm 0x94fbd90 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb445fc8e45153ce3 - Init COMPLETE +ip-26-0-168-34:70613:71046 [2] NCCL INFO comm 0x98dee90 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb445fc8e45153ce3 - Init COMPLETE +ip-26-0-168-34:70617:71043 [6] NCCL INFO comm 0x8ba9050 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb445fc8e45153ce3 - Init COMPLETE +ip-26-0-168-34:70616:71048 [5] NCCL INFO comm 0xa65d6d0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb445fc8e45153ce3 - Init COMPLETE +ip-26-0-168-34:70611:71042 [0] NCCL INFO comm 0x9843040 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb445fc8e45153ce3 - Init COMPLETE +ip-26-0-168-34:70614:71049 [3] NCCL INFO comm 0x9f34f60 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb445fc8e45153ce3 - Init COMPLETE +ip-26-0-168-34:70612:71045 [1] NCCL INFO comm 0x8b8abf0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb445fc8e45153ce3 - Init COMPLETE +ip-26-0-168-34:70618:71044 [7] NCCL INFO comm 0x98b7520 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb445fc8e45153ce3 - Init COMPLETE +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Connected all trees +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO NVLS comm 0xa361700 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25207:25646 [2] NCCL INFO Connected all trees +ip-26-0-162-180:25207:25646 [2] NCCL INFO NVLS comm 0x89fe720 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Connected all trees +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO NVLS comm 0x9a71010 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Connected all trees +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO NVLS comm 0x9a96820 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25211:25641 [6] NCCL INFO Connected all trees +ip-26-0-162-180:25211:25641 [6] NCCL INFO NVLS comm 0xa17a3b0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25206:25647 [1] NCCL INFO Connected all trees +ip-26-0-162-180:25206:25647 [1] NCCL INFO NVLS comm 0x9df86d0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25210:25644 [5] NCCL INFO Connected all trees +ip-26-0-162-180:25210:25644 [5] NCCL INFO NVLS comm 0x97bca90 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-162-180:25208:25642 [3] NCCL INFO Connected all trees +ip-26-0-162-180:25208:25642 [3] NCCL INFO NVLS comm 0x97f9da0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-162-180:25209:25645 [4] NCCL INFO Connected all trees +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO Connected NVLS tree +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-162-180:25209:25645 [4] NCCL INFO NVLS comm 0x9739ef0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-168-95:1736129:1736556 [6] NCCL INFO comm 0xa5359c0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x14693696db1c5513 - Init COMPLETE +ip-26-0-168-95:1736127:1736557 [4] NCCL INFO comm 0x8844760 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x14693696db1c5513 - Init COMPLETE +ip-26-0-168-95:1736125:1736558 [2] NCCL INFO comm 0x9a71010 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x14693696db1c5513 - Init COMPLETE +ip-26-0-168-95:1736123:1736553 [0] NCCL INFO comm 0x930a490 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x14693696db1c5513 - Init COMPLETE +ip-26-0-168-95:1736130:1736560 [7] NCCL INFO comm 0x8eff3f0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x14693696db1c5513 - Init COMPLETE +ip-26-0-168-95:1736126:1736555 [3] NCCL INFO comm 0xa361700 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x14693696db1c5513 - Init COMPLETE +ip-26-0-168-95:1736128:1736554 [5] NCCL INFO comm 0xa0ba010 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x14693696db1c5513 - Init COMPLETE +ip-26-0-168-95:1736124:1736559 [1] NCCL INFO comm 0x9a96820 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x14693696db1c5513 - Init COMPLETE +ip-26-0-162-180:25211:25641 [6] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25211:25641 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25211:25641 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-180:25211:25641 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-162-180:25212:25643 [7] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25212:25643 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25212:25643 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-180:25212:25643 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-162-180:25208:25642 [3] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25208:25642 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25208:25642 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-180:25208:25642 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-162-180:25207:25646 [2] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25210:25644 [5] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25207:25646 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25207:25646 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-180:25207:25646 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-162-180:25210:25644 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25210:25644 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-180:25210:25644 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-162-180:25209:25645 [4] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25209:25645 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25209:25645 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-180:25209:25645 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-162-180:25206:25647 [1] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25206:25647 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25206:25647 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-180:25206:25647 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-162-180:25205:25640 [0] NCCL INFO Connected NVLS tree +ip-26-0-162-180:25205:25640 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-162-180:25205:25640 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-162-180:25205:25640 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-162-180:25210:25644 [5] NCCL INFO comm 0x97bca90 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x5e3dea37485e3e97 - Init COMPLETE +ip-26-0-162-180:25212:25643 [7] NCCL INFO comm 0x9494a60 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5e3dea37485e3e97 - Init COMPLETE +ip-26-0-162-180:25208:25642 [3] NCCL INFO comm 0x97f9da0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x5e3dea37485e3e97 - Init COMPLETE +ip-26-0-162-180:25209:25645 [4] NCCL INFO comm 0x9739ef0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x5e3dea37485e3e97 - Init COMPLETE +ip-26-0-162-180:25211:25641 [6] NCCL INFO comm 0xa17a3b0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x5e3dea37485e3e97 - Init COMPLETE +ip-26-0-162-180:25206:25647 [1] NCCL INFO comm 0x9df86d0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x5e3dea37485e3e97 - Init COMPLETE +ip-26-0-162-180:25207:25646 [2] NCCL INFO comm 0x89fe720 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5e3dea37485e3e97 - Init COMPLETE +ip-26-0-162-180:25205:25640 [0] NCCL INFO comm 0x96bb550 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x5e3dea37485e3e97 - Init COMPLETE +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Config: +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Config(general=GeneralArgs(project='debug', +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: run='469G_dp4_tp4_pp2_acc64_mbs1_seq4096_zero1_tpmodeRED_vocab131k', +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: seed=42, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: step=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: consumed_train_samples=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: ignore_sanity_checks=True), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: parallelism=ParallelismArgs(dp=4, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pp=2, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tp=4, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pp_engine=, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tp_mode=, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tp_linear_async_communication=True, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: recompute_layer=False, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tp_recompute_allgather=True, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: expert_parallel_size=1), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: eos_token_id=0, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: hidden_act='silu', +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: hidden_size=16384, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: initializer_range=0.02, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: intermediate_size=53248, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: is_llama_config=True, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: max_position_embeddings=4096, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_attention_heads=128, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_hidden_layers=126, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_key_value_heads=128, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pad_token_id=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pretraining_tp=1, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rms_norm_eps=1e-05, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_scaling=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_theta=10000.0, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_interleaved=False, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tie_word_embeddings=False, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: use_cache=True, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: vocab_size=131072), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: init_method=RandomInit(std=0.02), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: dtype=torch.bfloat16, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: make_vocab_size_divisible_by=1, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: ddp_bucket_cap_mb=25), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tokenizer_revision=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tokenizer_max_length=None), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: checkpoint_interval=10000, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: save_initial_state=False, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: save_final_state=False, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: resume_checkpoint_path=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: checkpoints_path_is_shared_file_system=False), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: logging=LoggingArgs(log_level='info', +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: log_level_replica='info', +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: iteration_step_info_interval=1), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tokens=TokensArgs(sequence_length=4096, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: train_steps=100, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: micro_batch_size=1, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: batch_accumulation_per_replica=64, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: val_check_interval=100, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: limit_val_batches=0, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: limit_test_batches=0), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: adam_beta1=0.9, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: adam_beta2=0.95, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: torch_adam_is_fused=True, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: name='adamW'), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: zero_stage=1, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: weight_decay=0.01, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: clip_grad=1.0, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: accumulate_grad_in_fp32=True, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lr_warmup_steps=2, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lr_warmup_style='linear', +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lr_decay_style='cosine', +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lr_decay_steps=13, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lr_decay_starting_step=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: min_decay_lr=1e-05)), +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: start_training_step=1, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: data=DataArgs(dataset=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: seed=42, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_loading_workers=1))], +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: profiler=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: lighteval=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: s3_upload=None) +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Model Config: +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: LlamaConfig(bos_token_id=0, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: eos_token_id=0, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: hidden_act='silu', +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: hidden_size=16384, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: initializer_range=0.02, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: intermediate_size=53248, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: is_llama_config=True, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: max_position_embeddings=4096, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_attention_heads=128, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_hidden_layers=126, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: num_key_value_heads=128, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pad_token_id=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: pretraining_tp=1, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rms_norm_eps=1e-05, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_scaling=None, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_theta=10000.0, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: rope_interleaved=False, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: tie_word_embeddings=False, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: use_cache=True, +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: vocab_size=131072) +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Building model.. +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Initialize RoPE Theta = 10000.0 +12/28/2024 02:05:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-162-180]: Setting PP block ranks... +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 128.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.13 GiB is allocated by PyTorch, and 961.42 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 69.35 GiB is allocated by PyTorch, and 961.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 69.25 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.25 GiB is allocated by PyTorch, and 961.39 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 69.35 GiB is allocated by PyTorch, and 961.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 69.25 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + trainer = DistributedTrainer(config_file) File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + model = self._init_model( + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 499.94 MiB is free. Including non-PyTorch memory, this process has 78.83 GiB memory in use. Of the allocated memory 69.25 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 363.94 MiB is free. Including non-PyTorch memory, this process has 78.96 GiB memory in use. Of the allocated memory 69.25 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.down_proj = TensorParallelRowLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 347.94 MiB is free. Including non-PyTorch memory, this process has 78.98 GiB memory in use. Of the allocated memory 70.07 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 283.94 MiB is free. Including non-PyTorch memory, this process has 79.04 GiB memory in use. Of the allocated memory 69.35 GiB is allocated by PyTorch, and 961.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs))self.model = self.init_model() # Defines self.model + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 69.25 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 411.94 MiB is free. Including non-PyTorch memory, this process has 78.92 GiB memory in use. Of the allocated memory 69.25 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 69.25 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + self.qkv_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 69.75 GiB is allocated by PyTorch, and 961.42 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 51.94 MiB is free. Including non-PyTorch memory, this process has 79.27 GiB memory in use. Of the allocated memory 70.25 GiB is allocated by PyTorch, and 961.39 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 51.94 MiB is free. Including non-PyTorch memory, this process has 79.27 GiB memory in use. Of the allocated memory 70.25 GiB is allocated by PyTorch, and 961.39 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 331.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 69.35 GiB is allocated by PyTorch, and 961.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 147.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.07 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 35.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.25 GiB is allocated by PyTorch, and 961.39 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.down_proj = TensorParallelRowLinear( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + model = self._init_model( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 69.35 GiB is allocated by PyTorch, and 961.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.07 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 51.94 MiB is free. Including non-PyTorch memory, this process has 79.27 GiB memory in use. Of the allocated memory 70.25 GiB is allocated by PyTorch, and 961.39 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 69.25 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 147.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.07 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 69.35 GiB is allocated by PyTorch, and 961.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 51.94 MiB is free. Including non-PyTorch memory, this process has 79.27 GiB memory in use. Of the allocated memory 70.25 GiB is allocated by PyTorch, and 961.39 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.down_proj = TensorParallelRowLinear( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 69.35 GiB is allocated by PyTorch, and 961.45 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 147.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.07 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 147.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.07 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 832.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 251.94 MiB is free. Including non-PyTorch memory, this process has 79.07 GiB memory in use. Of the allocated memory 70.25 GiB is allocated by PyTorch, and 961.39 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.07 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 231, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 233, in __init__ + self.down_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 329, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 27.94 MiB is free. Including non-PyTorch memory, this process has 79.29 GiB memory in use. Of the allocated memory 70.07 GiB is allocated by PyTorch, and 1.03 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +ip-26-0-162-180:25205:25664 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25212:25622 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25210:25624 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25211:25621 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25206:25619 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25209:25625 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25205:25618 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25207:25620 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25212:25391 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25210:25390 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25211:25395 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25209:25392 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25206:25389 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25207:25393 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25205:25394 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-162-180:25205:25664 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25211:25621 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25210:25624 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25207:25620 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25206:25619 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25205:25618 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25212:25622 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25211:25395 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25210:25390 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25207:25393 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25206:25389 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25205:25394 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25212:25391 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-162-180:25205:25664 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-162-180:25211:25621 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-162-180:25207:25620 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-162-180:25206:25619 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-162-180:25205:25618 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-162-180:25211:25395 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-162-180:25207:25393 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-162-180:25206:25389 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-162-180:25205:25394 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-168-95:1736123:1736576 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-95:1736130:1736532 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-95:1736127:1736538 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-95:1736126:1736536 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-95:1736124:1736537 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-95:1736123:1736535 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-95:1736130:1736306 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-95:1736127:1736303 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-95:1736126:1736309 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-95:1736124:1736307 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-95:1736123:1736308 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-95:1736123:1736576 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-168-95:1736130:1736532 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-168-95:1736127:1736538 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-168-95:1736126:1736536 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-168-95:1736123:1736535 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-168-95:1736124:1736537 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-168-95:1736126:1736309 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-168-95:1736123:1736308 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-168-95:1736124:1736307 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-168-95:1736130:1736306 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-162-180:25205:25664 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-162-180:25211:25621 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-162-180:25205:25618 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-162-180:25211:25395 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-162-180:25205:25394 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-168-95:1736123:1736576 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-168-95:1736126:1736536 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-168-95:1736123:1736535 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-168-95:1736126:1736309 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-168-95:1736123:1736308 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-168-95:1736123:1736576 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-95:1736123:1736535 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-95:1736126:1736536 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-95:1736123:1736308 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-95:1736126:1736309 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70611:71060 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70618:71026 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70615:71025 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70612:71021 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70614:71027 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70613:71022 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70611:71023 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70616:71028 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70618:70794 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70615:70792 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70614:70797 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70613:70790 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70612:70795 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70611:70791 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70616:70796 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-168-34:70611:71060 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70616:71028 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70613:71022 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70611:71023 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70612:71021 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70614:71027 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70616:70796 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70613:70790 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70612:70795 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70611:70791 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70614:70797 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-168-34:70611:71060 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-168-34:70616:71028 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-168-34:70613:71022 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-168-34:70611:71023 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-168-34:70616:70796 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-168-34:70613:70790 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-168-34:70611:70791 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-168-34:70611:71060 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-168-34:70616:71028 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-168-34:70611:71023 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-168-34:70613:71022 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-168-34:70616:70796 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-168-34:70611:70791 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-168-34:70613:70790 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-167-217:138056:138508 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138063:138469 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138061:138465 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138062:138468 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138060:138472 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138059:138471 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138058:138466 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138056:138467 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138063:138240 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138062:138239 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138061:138241 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138060:138236 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138059:138243 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138058:138238 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138056:138242 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-168-34:70611:71060 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-168-34:70616:71028 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-168-34:70613:71022 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-168-34:70611:71023 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-168-34:70616:70796 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-168-34:70613:70790 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-168-34:70611:70791 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-167-217:138056:138508 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-167-217:138056:138467 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-167-217:138058:138466 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-167-217:138056:138242 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-167-217:138058:138238 [0] NCCL INFO [Service thread] Connection closed by localRank 3 +ip-26-0-167-217:138056:138508 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-167-217:138058:138466 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-167-217:138056:138467 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-167-217:138058:138238 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-167-217:138056:138242 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-167-217:138056:138508 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-167-217:138058:138466 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-167-217:138056:138467 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-167-217:138058:138238 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-167-217:138056:138242 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-167-217:138056:138508 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-167-217:138056:138467 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-167-217:138056:138242 [0] NCCL INFO [Service thread] Connection closed by localRank 6 +ip-26-0-167-217:138056:138508 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-167-217:138056:138467 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +ip-26-0-167-217:138056:138242 [0] NCCL INFO [Service thread] Connection closed by localRank 5 +[2024-12-28 02:06:08,419] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1736123) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-28 02:06:08,419] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 25212 closing signal SIGTERM +[2024-12-28 02:06:08,425] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 138058 closing signal SIGTERM +[2024-12-28 02:06:08,425] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 138060 closing signal SIGTERM +[2024-12-28 02:06:08,425] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 138061 closing signal SIGTERM +[2024-12-28 02:06:08,425] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 138062 closing signal SIGTERM +[2024-12-28 02:06:08,425] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 138063 closing signal SIGTERM +[2024-12-28 02:06:08,428] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 70612 closing signal SIGTERM +[2024-12-28 02:06:08,428] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 70613 closing signal SIGTERM +[2024-12-28 02:06:08,428] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 70614 closing signal SIGTERM +[2024-12-28 02:06:08,428] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 70615 closing signal SIGTERM +[2024-12-28 02:06:08,428] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 70616 closing signal SIGTERM +[2024-12-28 02:06:08,428] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 70618 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-28_02:06:08 + host : ip-26-0-168-95.ec2.internal + rank : 25 (local_rank: 1) + exitcode : 1 (pid: 1736124) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-28_02:06:08 + host : ip-26-0-168-95.ec2.internal + rank : 26 (local_rank: 2) + exitcode : 1 (pid: 1736125) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-28_02:06:08 + host : ip-26-0-168-95.ec2.internal + rank : 27 (local_rank: 3) + exitcode : 1 (pid: 1736126) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-28_02:06:08 + host : ip-26-0-168-95.ec2.internal + rank : 28 (local_rank: 4) + exitcode : 1 (pid: 1736127) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-28_02:06:08 + host : ip-26-0-168-95.ec2.internal + rank : 29 (local_rank: 5) + exitcode : 1 (pid: 1736128) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-28_02:06:08 + host : ip-26-0-168-95.ec2.internal + rank : 30 (local_rank: 6) + exitcode : 1 (pid: 1736129) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2024-12-28_02:06:08 + host : ip-26-0-168-95.ec2.internal + rank : 31 (local_rank: 7) + exitcode : 1 (pid: 1736130) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-28_02:06:08 + host : ip-26-0-168-95.ec2.internal + rank : 24 (local_rank: 0) + exitcode : 1 (pid: 1736123) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2024-12-28 02:06:08,583] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 25205) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-28_02:06:08 + host : ip-26-0-162-180.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 25206) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-28_02:06:08 + host : ip-26-0-162-180.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 25207) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2024-12-28_02:06:08 + host : ip-26-0-162-180.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 25208) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2024-12-28_02:06:08 + host : ip-26-0-162-180.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 25209) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2024-12-28_02:06:08 + host : ip-26-0-162-180.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 25210) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2024-12-28_02:06:08 + host : ip-26-0-162-180.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 25211) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-28_02:06:08 + host : ip-26-0-162-180.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 25205) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-168-95: task 3: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13848535.0 +slurmstepd: error: *** STEP 13848535.0 ON ip-26-0-162-180 CANCELLED AT 2024-12-28T02:06:08 *** +[2024-12-28 02:06:08,746] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-28 02:06:08,746] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 70612 closing signal SIGTERM +[2024-12-28 02:06:08,746] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 70613 closing signal SIGTERM +[2024-12-28 02:06:08,746] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 70614 closing signal SIGTERM +[2024-12-28 02:06:08,745] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-28 02:06:08,746] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 70615 closing signal SIGTERM +[2024-12-28 02:06:08,746] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 70616 closing signal SIGTERM +[2024-12-28 02:06:08,745] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 138058 closing signal SIGTERM +[2024-12-28 02:06:08,746] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 138060 closing signal SIGTERM +[2024-12-28 02:06:08,746] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 138061 closing signal SIGTERM +[2024-12-28 02:06:08,746] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 138062 closing signal SIGTERM +[2024-12-28 02:06:08,746] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 138063 closing signal SIGTERM +srun: error: ip-26-0-162-180: task 0: Terminated +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 137985 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 70538 got signal: 15 +srun: error: ip-26-0-167-217: task 1: Exited with exit code 1 +srun: error: ip-26-0-168-34: task 2: Exited with exit code 1 +srun: Force Terminated StepId=13848535.0 diff --git a/logs/13848550-bench_8.86G_dp8_tp1_pp4_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13848550-bench_8.86G_dp8_tp1_pp4_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..66fc034c0ead399336bc113a92e4a7f085d462ee --- /dev/null +++ b/logs/13848550-bench_8.86G_dp8_tp1_pp4_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,14801 @@ ++ '[' -z 13848550 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-172-[73,116,142,147]' ++ export 'NODELIST=ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147' ++ NODELIST='ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-172-[73,116,142,147]' ++ export MASTER_NODE=ip-26-0-172-73 ++ MASTER_NODE=ip-26-0-172-73 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=INFO ++ NCCL_DEBUG=INFO ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-172-73' +Master node: ip-26-0-172-73 ++ echo 'All nodes: ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147' +All nodes: ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 ++ echo 'World size: 32' +World size: 32 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=13848550 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-172-73:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_8.86G_dp8_tp1_pp4_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-28 02:07:46,174] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 02:07:46,180] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 02:07:46,194] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 02:07:46,175] torch.distributed.run: [WARNING] +[2024-12-28 02:07:46,175] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:07:46,175] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 02:07:46,175] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:07:46,180] torch.distributed.run: [WARNING] +[2024-12-28 02:07:46,180] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:07:46,180] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 02:07:46,180] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:07:46,194] torch.distributed.run: [WARNING] +[2024-12-28 02:07:46,194] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:07:46,194] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 02:07:46,194] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:07:46,371] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-28 02:07:46,371] torch.distributed.run: [WARNING] +[2024-12-28 02:07:46,371] torch.distributed.run: [WARNING] ***************************************** +[2024-12-28 02:07:46,371] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-28 02:07:46,371] torch.distributed.run: [WARNING] ***************************************** +ip-26-0-172-116:3888248:3888248 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-116:3888248:3888248 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.172.116<0> +ip-26-0-172-116:3888248:3888248 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-116:3888248:3888248 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-116:3888248:3888248 [0] NCCL INFO cudaDriverVersion 12020 +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-116:3888250:3888250 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-116:3888252:3888252 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-116:3888250:3888250 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-116:3888252:3888252 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-116:3888250:3888250 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.172.116<0> +ip-26-0-172-116:3888252:3888252 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.172.116<0> +ip-26-0-172-116:3888252:3888252 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-116:3888250:3888250 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-116:3888252:3888252 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-116:3888250:3888250 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-73:2574285:2574285 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-73:2574287:2574287 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-73:2574282:2574282 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-73:2574289:2574289 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-73:2574288:2574288 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-73:2574286:2574286 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-73:2574285:2574285 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-73:2574287:2574287 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-73:2574282:2574282 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-73:2574289:2574289 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-73:2574288:2574288 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-73:2574286:2574286 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-73:2574283:2574283 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-73:2574284:2574284 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-73:2574283:2574283 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-73:2574284:2574284 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-142:114199:114199 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-142:114196:114196 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-142:114193:114193 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-147:2713680:2713680 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-142:114194:114194 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-142:114198:114198 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-142:114192:114192 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-142:114197:114197 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-142:114195:114195 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-147:2713676:2713676 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-142:114199:114199 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-142:114194:114194 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-116:3888254:3888254 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-142:114198:114198 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-142:114196:114196 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-142:114192:114192 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-142:114197:114197 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-147:2713680:2713680 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-142:114195:114195 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-142:114193:114193 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-147:2713676:2713676 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-116:3888254:3888254 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-73:2574285:2574285 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.172.73<0> +ip-26-0-172-116:3888254:3888254 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.172.116<0> +ip-26-0-172-73:2574289:2574289 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.172.73<0> +ip-26-0-172-73:2574282:2574282 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.172.73<0> +ip-26-0-172-73:2574287:2574287 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.172.73<0> +ip-26-0-172-73:2574288:2574288 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.172.73<0> +ip-26-0-172-73:2574286:2574286 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.172.73<0> +ip-26-0-172-73:2574284:2574284 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.172.73<0> +ip-26-0-172-73:2574283:2574283 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.172.73<0> +ip-26-0-172-116:3888254:3888254 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-116:3888254:3888254 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-73:2574285:2574285 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-73:2574285:2574285 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713680:2713680 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.172.147<0> +ip-26-0-172-147:2713676:2713676 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.172.147<0> +ip-26-0-172-73:2574288:2574288 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-73:2574284:2574284 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-73:2574283:2574283 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-73:2574288:2574288 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-73:2574284:2574284 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-73:2574283:2574283 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-73:2574282:2574282 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-73:2574289:2574289 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-73:2574289:2574289 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-73:2574282:2574282 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-73:2574286:2574286 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-73:2574286:2574286 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713680:2713680 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-147:2713680:2713680 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713674:2713674 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-147:2713676:2713676 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-147:2713676:2713676 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-73:2574287:2574287 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-73:2574287:2574287 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713675:2713675 [2] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-147:2713673:2713673 [0] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-147:2713674:2713674 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-147:2713679:2713679 [6] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-147:2713675:2713675 [2] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-116:3888249:3888249 [1] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-147:2713673:2713673 [0] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-147:2713678:2713678 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-147:2713679:2713679 [6] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-147:2713678:2713678 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-116:3888253:3888253 [5] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-116:3888249:3888249 [1] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-147:2713677:2713677 [4] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-116:3888253:3888253 [5] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-116:3888255:3888255 [7] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-147:2713677:2713677 [4] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-116:3888255:3888255 [7] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-142:114194:114194 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.172.142<0> +ip-26-0-172-142:114199:114199 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.172.142<0> +ip-26-0-172-142:114198:114198 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.172.142<0> +ip-26-0-172-116:3888251:3888251 [3] NCCL INFO cudaDriverVersion 12020 +ip-26-0-172-142:114193:114193 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.172.142<0> +ip-26-0-172-142:114196:114196 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.172.142<0> +ip-26-0-172-142:114192:114192 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.172.142<0> +ip-26-0-172-142:114197:114197 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.172.142<0> +ip-26-0-172-142:114195:114195 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.172.142<0> +ip-26-0-172-116:3888251:3888251 [3] NCCL INFO NCCL_SOCKET_IFNAME set by environment to enp +ip-26-0-172-116:3888249:3888249 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.172.116<0> +ip-26-0-172-142:114198:114198 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-142:114194:114194 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-142:114193:114193 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-142:114193:114193 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-142:114194:114194 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-142:114198:114198 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-116:3888253:3888253 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.172.116<0> +ip-26-0-172-142:114199:114199 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-142:114199:114199 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-142:114195:114195 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-142:114192:114192 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-142:114197:114197 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-142:114195:114195 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-142:114192:114192 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-142:114197:114197 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713674:2713674 [1] NCCL INFO Bootstrap : Using enp71s0:26.0.172.147<0> +ip-26-0-172-142:114196:114196 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-142:114196:114196 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-116:3888255:3888255 [7] NCCL INFO Bootstrap : Using enp71s0:26.0.172.116<0> +ip-26-0-172-116:3888251:3888251 [3] NCCL INFO Bootstrap : Using enp71s0:26.0.172.116<0> +ip-26-0-172-116:3888249:3888249 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-116:3888249:3888249 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-116:3888253:3888253 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-116:3888253:3888253 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713675:2713675 [2] NCCL INFO Bootstrap : Using enp71s0:26.0.172.147<0> +ip-26-0-172-147:2713673:2713673 [0] NCCL INFO Bootstrap : Using enp71s0:26.0.172.147<0> +ip-26-0-172-147:2713674:2713674 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-147:2713674:2713674 [1] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-116:3888255:3888255 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-116:3888255:3888255 [7] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713679:2713679 [6] NCCL INFO Bootstrap : Using enp71s0:26.0.172.147<0> +ip-26-0-172-147:2713678:2713678 [5] NCCL INFO Bootstrap : Using enp71s0:26.0.172.147<0> +ip-26-0-172-116:3888251:3888251 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-116:3888251:3888251 [3] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713677:2713677 [4] NCCL INFO Bootstrap : Using enp71s0:26.0.172.147<0> +ip-26-0-172-147:2713673:2713673 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-147:2713673:2713673 [0] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713675:2713675 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-147:2713675:2713675 [2] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713678:2713678 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-147:2713679:2713679 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-147:2713678:2713678 [5] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713679:2713679 [6] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713677:2713677 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin_v6 symbol. +ip-26-0-172-147:2713677:2713677 [4] NCCL INFO NET/Plugin: Failed to find ncclCollNetPlugin symbol (v4 or v5). +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Initializing aws-ofi-nccl 1.13.2-aws +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Using Libfabric version 1.22 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Using CUDA driver version 12020 with runtime 12020 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Configuring AWS-specific options +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Setting NCCL_NET_FORCE_FLUSH=0 for Hopper GPUs +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Setting NCCL_NVLSTREE_MAX_CHUNKSIZE to 512KiB +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Setting NCCL_NVLS_CHUNKSIZE to 512KiB +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Internode latency set at 75.0 us +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Using transport protocol RDMA (platform set) +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Selected Provider is efa (found 32 nics) +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-142:114197:114357 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-142:114197:114357 [5] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/253 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 0 device #0 0000:52:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 0 device #1 0000:51:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 0 device #2 0000:50:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 0 device #3 0000:4f:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 1 device #0 0000:63:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 1 device #1 0000:62:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 1 device #2 0000:61:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 1 device #3 0000:60:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 2 device #0 0000:74:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 2 device #1 0000:73:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 2 device #2 0000:72:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 2 device #3 0000:71:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 3 device #0 0000:85:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 3 device #1 0000:84:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 3 device #2 0000:83:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 3 device #3 0000:82:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 4 device #0 0000:96:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 4 device #1 0000:95:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 4 device #2 0000:94:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 4 device #3 0000:93:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 5 device #0 0000:a7:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 5 device #1 0000:a6:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 5 device #2 0000:a5:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 5 device #3 0000:a4:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 6 device #0 0000:b8:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 6 device #1 0000:b7:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 6 device #2 0000:b6:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 6 device #3 0000:b5:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 7 device #0 0000:c9:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 7 device #1 0000:c8:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 7 device #2 0000:c7:00.0 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI NIC group 7 device #3 0000:c6:00.0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Setting NCCL_TOPO_FILE environment variable to /proc/self/fd/216 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Creating one domain per process +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Using network Libfabric +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-142:114195:114355 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-142:114195:114355 [3] NCCL INFO Using network Libfabric +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-142:114199:114356 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-142:114199:114356 [7] NCCL INFO Using network Libfabric +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-142:114193:114354 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-142:114193:114354 [1] NCCL INFO Using network Libfabric +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-142:114192:114352 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-142:114192:114352 [0] NCCL INFO Using network Libfabric +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-142:114194:114353 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-142:114194:114353 [2] NCCL INFO Using network Libfabric +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-142:114196:114358 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-142:114196:114358 [4] NCCL INFO Using network Libfabric +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-142:114198:114351 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-142:114198:114351 [6] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Support for global registrations: true +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NET/OFI Support for DMA-BUF registrations: false +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO comm 0x9ab1f90 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO comm 0x8815b80 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO comm 0x9bf8090 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO comm 0x8cc3430 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO comm 0x8691790 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO comm 0x8786fe0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO comm 0xa03fb30 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO comm 0xa259700 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO comm 0x93f4e10 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO comm 0x9bab5e0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO comm 0x9523840 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO comm 0x939e880 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO comm 0xa341fc0 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO comm 0x919d720 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO comm 0x92753a0 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-142:114199:114356 [7] NCCL INFO comm 0x94ed670 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO comm 0x9779bd0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO comm 0x92f7c20 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO comm 0xa37bf90 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO comm 0x9e18770 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO comm 0xa05ea30 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO comm 0x991d9c0 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO comm 0x864c570 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO comm 0x97bc350 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO comm 0xa351520 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-142:114192:114352 [0] NCCL INFO comm 0x99232a0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-142:114196:114358 [4] NCCL INFO comm 0x8953d90 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-142:114197:114357 [5] NCCL INFO comm 0xa2e15f0 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-142:114195:114355 [3] NCCL INFO comm 0x987a8c0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-142:114193:114354 [1] NCCL INFO comm 0x9968620 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-142:114194:114353 [2] NCCL INFO comm 0x94694e0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-142:114198:114351 [6] NCCL INFO comm 0x9e6bdf0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb0728d2e94633fff - Init START +ip-26-0-172-142:114199:114356 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114195:114355 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114194:114353 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114193:114354 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-142:114198:114351 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114198:114351 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-142:114195:114355 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-142:114195:114355 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-142:114193:114354 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-142:114193:114354 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-142:114194:114353 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-142:114194:114353 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-142:114196:114358 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114196:114358 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-142:114192:114352 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-142:114192:114352 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114197:114357 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-142:114199:114356 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114199:114356 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/17/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] 24/-1/-1->31->30 [2] 24/-1/-1->31->30 [3] 24/-1/-1->31->30 [4] 24/-1/-1->31->30 [5] 24/-1/-1->31->30 [6] 24/-1/-1->31->30 [7] 24/-1/-1->31->23 [8] -1/-1/-1->31->30 [9] 24/-1/-1->31->30 [10] 24/-1/-1->31->30 [11] 24/-1/-1->31->30 [12] 24/-1/-1->31->30 [13] 24/-1/-1->31->30 [14] 24/-1/-1->31->30 [15] 24/15/-1->31->-1 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Trees [0] 30/-1/-1->29->28 [1] 30/-1/-1->29->28 [2] 30/-1/-1->29->28 [3] 30/-1/-1->29->28 [4] 30/-1/-1->29->28 [5] 30/-1/-1->29->21 [6] -1/-1/-1->29->28 [7] 30/-1/-1->29->28 [8] 30/-1/-1->29->28 [9] 30/-1/-1->29->28 [10] 30/-1/-1->29->28 [11] 30/-1/-1->29->28 [12] 30/-1/-1->29->28 [13] 30/13/-1->29->-1 [14] -1/-1/-1->29->28 [15] 30/-1/-1->29->28 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/18/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Trees [0] 31/-1/-1->30->29 [1] 31/-1/-1->30->29 [2] 31/-1/-1->30->29 [3] 31/-1/-1->30->29 [4] 31/-1/-1->30->29 [5] 31/-1/-1->30->29 [6] 31/-1/-1->30->22 [7] -1/-1/-1->30->29 [8] 31/-1/-1->30->29 [9] 31/-1/-1->30->29 [10] 31/-1/-1->30->29 [11] 31/-1/-1->30->29 [12] 31/-1/-1->30->29 [13] 31/-1/-1->30->29 [14] 31/14/-1->30->-1 [15] -1/-1/-1->30->29 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 01/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Trees [0] 28/-1/-1->27->26 [1] 28/-1/-1->27->26 [2] 28/-1/-1->27->26 [3] 28/-1/-1->27->19 [4] -1/-1/-1->27->26 [5] 28/-1/-1->27->26 [6] 28/-1/-1->27->26 [7] 28/-1/-1->27->26 [8] 28/-1/-1->27->26 [9] 28/-1/-1->27->26 [10] 28/-1/-1->27->26 [11] 28/11/-1->27->-1 [12] -1/-1/-1->27->26 [13] 28/-1/-1->27->26 [14] 28/-1/-1->27->26 [15] 28/-1/-1->27->26 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Trees [0] 26/-1/-1->25->24 [1] 26/-1/-1->25->17 [2] -1/-1/-1->25->24 [3] 26/-1/-1->25->24 [4] 26/-1/-1->25->24 [5] 26/-1/-1->25->24 [6] 26/-1/-1->25->24 [7] 26/-1/-1->25->24 [8] 26/-1/-1->25->24 [9] 26/9/-1->25->-1 [10] -1/-1/-1->25->24 [11] 26/-1/-1->25->24 [12] 26/-1/-1->25->24 [13] 26/-1/-1->25->24 [14] 26/-1/-1->25->24 [15] 26/-1/-1->25->24 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 03/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] 16/-1/-1->23->22 [2] 16/-1/-1->23->22 [3] 16/-1/-1->23->22 [4] 16/-1/-1->23->22 [5] 16/-1/-1->23->22 [6] 16/-1/-1->23->22 [7] 16/15/31->23->7 [8] -1/-1/-1->23->22 [9] 16/-1/-1->23->22 [10] 16/-1/-1->23->22 [11] 16/-1/-1->23->22 [12] 16/-1/-1->23->22 [13] 16/-1/-1->23->22 [14] 16/-1/-1->23->22 [15] 16/-1/-1->23->15 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Trees [0] 23/-1/-1->22->21 [1] 23/-1/-1->22->21 [2] 23/-1/-1->22->21 [3] 23/-1/-1->22->21 [4] 23/-1/-1->22->21 [5] 23/-1/-1->22->21 [6] 23/14/30->22->6 [7] -1/-1/-1->22->21 [8] 23/-1/-1->22->21 [9] 23/-1/-1->22->21 [10] 23/-1/-1->22->21 [11] 23/-1/-1->22->21 [12] 23/-1/-1->22->21 [13] 23/-1/-1->22->21 [14] 23/-1/-1->22->14 [15] -1/-1/-1->22->21 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Trees [0] 29/-1/-1->28->27 [1] 29/-1/-1->28->27 [2] 29/-1/-1->28->27 [3] 29/-1/-1->28->27 [4] 29/-1/-1->28->20 [5] -1/-1/-1->28->27 [6] 29/-1/-1->28->27 [7] 29/-1/-1->28->27 [8] 29/-1/-1->28->27 [9] 29/-1/-1->28->27 [10] 29/-1/-1->28->27 [11] 29/-1/-1->28->27 [12] 29/12/-1->28->-1 [13] -1/-1/-1->28->27 [14] 29/-1/-1->28->27 [15] 29/-1/-1->28->27 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 04/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 05/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-142:114192:114352 [0] NCCL INFO Trees [0] 9/-1/-1->8->16 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/16/0->8->24 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 +ip-26-0-172-142:114192:114352 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Trees [0] 25/-1/-1->24->16 [1] -1/-1/-1->24->31 [2] 25/-1/-1->24->31 [3] 25/-1/-1->24->31 [4] 25/-1/-1->24->31 [5] 25/-1/-1->24->31 [6] 25/-1/-1->24->31 [7] 25/-1/-1->24->31 [8] 25/8/-1->24->-1 [9] -1/-1/-1->24->31 [10] 25/-1/-1->24->31 [11] 25/-1/-1->24->31 [12] 25/-1/-1->24->31 [13] 25/-1/-1->24->31 [14] 25/-1/-1->24->31 [15] 25/-1/-1->24->31 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 06/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 07/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Trees [0] 22/-1/-1->21->20 [1] 22/-1/-1->21->20 [2] 22/-1/-1->21->20 [3] 22/-1/-1->21->20 [4] 22/-1/-1->21->20 [5] 22/13/29->21->5 [6] -1/-1/-1->21->20 [7] 22/-1/-1->21->20 [8] 22/-1/-1->21->20 [9] 22/-1/-1->21->20 [10] 22/-1/-1->21->20 [11] 22/-1/-1->21->20 [12] 22/-1/-1->21->20 [13] 22/-1/-1->21->13 [14] -1/-1/-1->21->20 [15] 22/-1/-1->21->20 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 09/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-172-142:114199:114356 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->23 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/23/7->15->31 +ip-26-0-172-142:114199:114356 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Trees [0] 21/-1/-1->20->19 [1] 21/-1/-1->20->19 [2] 21/-1/-1->20->19 [3] 21/-1/-1->20->19 [4] 21/12/28->20->4 [5] -1/-1/-1->20->19 [6] 21/-1/-1->20->19 [7] 21/-1/-1->20->19 [8] 21/-1/-1->20->19 [9] 21/-1/-1->20->19 [10] 21/-1/-1->20->19 [11] 21/-1/-1->20->19 [12] 21/-1/-1->20->12 [13] -1/-1/-1->20->19 [14] 21/-1/-1->20->19 [15] 21/-1/-1->20->19 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Trees [0] 19/-1/-1->18->17 [1] 19/-1/-1->18->17 [2] 19/10/26->18->2 [3] -1/-1/-1->18->17 [4] 19/-1/-1->18->17 [5] 19/-1/-1->18->17 [6] 19/-1/-1->18->17 [7] 19/-1/-1->18->17 [8] 19/-1/-1->18->17 [9] 19/-1/-1->18->17 [10] 19/-1/-1->18->10 [11] -1/-1/-1->18->17 [12] 19/-1/-1->18->17 [13] 19/-1/-1->18->17 [14] 19/-1/-1->18->17 [15] 19/-1/-1->18->17 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114198:114351 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->22 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/22/6->14->30 [15] -1/-1/-1->14->13 +ip-26-0-172-142:114198:114351 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114197:114357 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->21 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/21/5->13->29 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Trees [0] 20/-1/-1->19->18 [1] 20/-1/-1->19->18 [2] 20/-1/-1->19->18 [3] 20/11/27->19->3 [4] -1/-1/-1->19->18 [5] 20/-1/-1->19->18 [6] 20/-1/-1->19->18 [7] 20/-1/-1->19->18 [8] 20/-1/-1->19->18 [9] 20/-1/-1->19->18 [10] 20/-1/-1->19->18 [11] 20/-1/-1->19->11 [12] -1/-1/-1->19->18 [13] 20/-1/-1->19->18 [14] 20/-1/-1->19->18 [15] 20/-1/-1->19->18 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Trees [0] 18/-1/-1->17->16 [1] 18/9/25->17->1 [2] -1/-1/-1->17->16 [3] 18/-1/-1->17->16 [4] 18/-1/-1->17->16 [5] 18/-1/-1->17->16 [6] 18/-1/-1->17->16 [7] 18/-1/-1->17->16 [8] 18/-1/-1->17->16 [9] 18/-1/-1->17->9 [10] -1/-1/-1->17->16 [11] 18/-1/-1->17->16 [12] 18/-1/-1->17->16 [13] 18/-1/-1->17->16 [14] 18/-1/-1->17->16 [15] 18/-1/-1->17->16 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Trees [0] 27/-1/-1->26->25 [1] 27/-1/-1->26->25 [2] 27/-1/-1->26->18 [3] -1/-1/-1->26->25 [4] 27/-1/-1->26->25 [5] 27/-1/-1->26->25 [6] 27/-1/-1->26->25 [7] 27/-1/-1->26->25 [8] 27/-1/-1->26->25 [9] 27/-1/-1->26->25 [10] 27/10/-1->26->-1 [11] -1/-1/-1->26->25 [12] 27/-1/-1->26->25 [13] 27/-1/-1->26->25 [14] 27/-1/-1->26->25 [15] 27/-1/-1->26->25 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114197:114357 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114196:114358 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->20 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/20/4->12->28 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 +ip-26-0-172-142:114196:114358 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 11/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-142:114193:114354 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->17 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/17/1->9->25 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 +ip-26-0-172-142:114193:114354 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114195:114355 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->19 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/19/3->11->27 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 +ip-26-0-172-142:114195:114355 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Trees [0] 17/8/24->16->0 [1] -1/-1/-1->16->23 [2] 17/-1/-1->16->23 [3] 17/-1/-1->16->23 [4] 17/-1/-1->16->23 [5] 17/-1/-1->16->23 [6] 17/-1/-1->16->23 [7] 17/-1/-1->16->23 [8] 17/-1/-1->16->8 [9] -1/-1/-1->16->23 [10] 17/-1/-1->16->23 [11] 17/-1/-1->16->23 [12] 17/-1/-1->16->23 [13] 17/-1/-1->16->23 [14] 17/-1/-1->16->23 [15] 17/-1/-1->16->23 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 12/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-142:114194:114353 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->18 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/18/2->10->26 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 +ip-26-0-172-142:114194:114353 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/19/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 13/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 14/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 15/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Trees [0] 1/16/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/21/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/23/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/22/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/20/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114192:114352 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 08/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 08/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 08/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 08/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 08/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 08/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 08/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 08/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 08/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 08/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 08/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 08/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 05/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 06/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 05/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 05/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-142:114195:114355 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-142:114194:114353 [2] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 13/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 05/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 06/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 06/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 14/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 13/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 06/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 13/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 14/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 14/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 13/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 14/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 05/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 13/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 08/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 08/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 05/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 06/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 13/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 14/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 05/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 13/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 07/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 15/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 06/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 06/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 06/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 14/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 14/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 14/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 07/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 07/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 07/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 15/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 15/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 15/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 01/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 01/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 02/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 03/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 02/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 03/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 04/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 04/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 09/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 09/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 10/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 10/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 11/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 11/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 12/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 12/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 09/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 06/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 05/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 05/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 14/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 06/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 13/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO NCCL_NET_FORCE_FLUSH set by environment to 0. +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 14/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 06/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 06/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 05/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 05/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 13/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 14/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 13/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 01/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 14/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 01/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 13/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 04/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 05/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 06/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 04/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 07/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 05/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 09/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 06/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 10/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 04/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 07/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 11/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 10/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 12/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 11/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 13/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 12/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 14/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 05/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 15/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 13/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 04/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 14/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 06/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 05/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 07/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 06/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 15/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 10/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 07/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 11/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 09/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 10/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 12/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 10/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 13/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 11/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 14/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 12/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 15/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 13/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 14/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 15/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 07/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 15/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 07/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 07/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 15/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 15/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 02/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 02/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 03/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 04/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 03/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 05/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 09/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 04/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 05/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 10/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 09/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 12/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 11/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 10/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 11/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 01/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 11/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 01/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 12/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 12/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 02/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 13/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 02/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 13/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 04/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 09/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 04/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 10/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 12/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 09/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 10/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 12/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 03/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 03/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 04/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 04/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 05/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 05/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 06/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 04/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 07/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 06/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 01/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 07/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 04/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 07/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 09/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 09/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 07/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 09/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 11/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 09/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 10/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 12/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 11/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 06/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 10/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 11/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 13/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 12/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 07/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 01/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 13/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 14/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 14/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 15/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 15/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 12/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 06/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 15/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 11/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 07/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 09/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 12/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 09/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 10/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 15/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 10/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 11/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 11/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 14/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 15/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 14/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 15/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Connected all rings +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Connected all rings +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Connected all rings +ip-26-0-172-142:114193:114354 [1] NCCL INFO Connected all rings +ip-26-0-172-142:114198:114351 [6] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Connected all rings +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Connected all rings +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Connected all rings +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Connected all rings +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 05/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Connected all rings +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Connected all rings +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Connected all rings +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 04/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 05/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 06/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 07/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 10/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 11/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 12/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 13/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 14/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 15/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 04/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 05/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 06/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 07/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 09/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 11/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 12/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 13/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Connected all rings +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 14/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Connected all rings +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 15/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Connected all rings +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Connected all rings +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Connected all rings +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 04/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 05/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 06/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 07/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 09/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Connected all rings +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Connected all rings +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 11/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 04/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 05/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Connected all rings +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 12/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Connected all rings +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 06/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 13/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 07/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 14/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 09/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 15/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Connected all rings +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 10/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 05/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Connected all rings +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 12/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Connected all rings +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 04/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 06/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 04/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 05/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 13/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 04/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 05/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 07/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 06/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 06/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 14/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 13/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 09/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 05/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 04/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 07/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 14/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 10/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 15/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 06/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 06/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 05/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 09/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 11/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 07/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Connected all rings +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 07/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 10/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 12/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 09/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 12/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 13/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Connected all rings +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 06/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 09/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 10/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 10/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 13/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 11/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 11/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 14/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 04/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 14/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 13/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 12/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 05/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 15/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 14/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 07/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 13/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 15/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 14/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 09/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 05/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 04/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 10/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 06/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 05/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 11/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 07/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 10/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 12/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 07/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 09/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 11/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 09/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 10/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 12/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 10/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 13/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 15/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 11/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 15/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 13/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 11/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 13/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 04/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 14/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 12/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 14/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 06/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 15/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 13/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 04/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 06/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 15/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 07/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 07/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 15/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 05/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 09/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 09/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 10/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 11/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 06/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 10/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 12/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 07/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 11/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 14/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 15/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 13/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 12/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 14/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 14/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 15/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 15/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 00/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 08/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 09/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 01/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 01/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 02/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 08/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 02/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 03/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 03/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 00/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 04/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 05/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 05/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 00/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 00/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 04/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 00/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 05/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 00/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 05/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 08/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 04/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 03/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 08/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 04/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 06/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 08/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 03/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 06/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 09/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 08/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 13/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 05/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 08/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 07/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 13/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 05/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 12/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 07/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 12/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 06/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 06/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 09/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 07/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 09/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 07/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 08/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 10/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 10/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 11/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 08/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 11/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 11/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 11/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 13/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 12/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 13/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 02/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 13/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 12/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 14/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 02/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 08/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 14/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 14/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 13/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 15/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 08/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 08/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 10/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 14/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 15/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 15/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 10/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 15/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 06/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 06/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 08/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 08/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 14/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 14/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Connected all trees +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NVLS comm 0x92f7c20 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Connected all trees +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NVLS comm 0x919d720 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Connected all trees +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NVLS comm 0x9bf8090 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Connected all trees +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NVLS comm 0x9ab1f90 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Connected all trees +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NVLS comm 0x8786fe0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Connected all trees +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NVLS comm 0xa341fc0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Connected all trees +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Connected all trees +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NVLS comm 0xa37bf90 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NVLS comm 0x8815b80 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Connected all trees +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NVLS comm 0xa259700 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Connected all trees +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NVLS comm 0x97bc350 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Connected all trees +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NVLS comm 0x8cc3430 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Connected all trees +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NVLS comm 0x92753a0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Connected all trees +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NVLS comm 0x8691790 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114194:114353 [2] NCCL INFO Connected all trees +ip-26-0-172-142:114194:114353 [2] NCCL INFO NVLS comm 0x94694e0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Connected all trees +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Connected all trees +ip-26-0-172-142:114197:114357 [5] NCCL INFO Connected all trees +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NVLS comm 0xa03fb30 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NVLS comm 0x9779bd0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114197:114357 [5] NCCL INFO NVLS comm 0xa2e15f0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114196:114358 [4] NCCL INFO Connected all trees +ip-26-0-172-142:114196:114358 [4] NCCL INFO NVLS comm 0x8953d90 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114193:114354 [1] NCCL INFO Connected all trees +ip-26-0-172-142:114193:114354 [1] NCCL INFO NVLS comm 0x9968620 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Connected all trees +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NVLS comm 0x9e18770 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114195:114355 [3] NCCL INFO Connected all trees +ip-26-0-172-142:114195:114355 [3] NCCL INFO NVLS comm 0x987a8c0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114199:114356 [7] NCCL INFO Connected all trees +ip-26-0-172-142:114199:114356 [7] NCCL INFO NVLS comm 0x94ed670 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114192:114352 [0] NCCL INFO Connected all trees +ip-26-0-172-142:114192:114352 [0] NCCL INFO NVLS comm 0x99232a0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Connected all trees +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NVLS comm 0x9523840 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Connected all trees +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NVLS comm 0x991d9c0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Connected all trees +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NVLS comm 0xa351520 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Connected all trees +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NVLS comm 0x9bab5e0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Connected all trees +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NVLS comm 0xa05ea30 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Connected all trees +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Connected all trees +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NVLS comm 0x93f4e10 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NVLS comm 0x864c570 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114198:114351 [6] NCCL INFO Connected all trees +ip-26-0-172-142:114198:114351 [6] NCCL INFO NVLS comm 0x9e6bdf0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Connected all trees +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NVLS comm 0x939e880 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114356 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114351 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114355 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114354 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114352 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114357 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114353 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114358 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888250:3888410 [2] NCCL INFO comm 0x92753a0 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-116:3888251:3888413 [3] NCCL INFO comm 0x97bc350 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-116:3888254:3888412 [6] NCCL INFO comm 0xa37bf90 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-116:3888252:3888411 [4] NCCL INFO comm 0x9e18770 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-116:3888255:3888415 [7] NCCL INFO comm 0x92f7c20 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-116:3888249:3888416 [1] NCCL INFO comm 0x919d720 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-116:3888253:3888414 [5] NCCL INFO comm 0x9779bd0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-116:3888248:3888409 [0] NCCL INFO comm 0xa341fc0 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114194:114353 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114194:114353 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114194:114353 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114194:114353 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114198:114351 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114198:114351 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114198:114351 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114198:114351 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114196:114358 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114196:114358 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114196:114358 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114196:114358 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114199:114356 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114199:114356 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114199:114356 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114199:114356 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114193:114354 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114193:114354 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114193:114354 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114193:114354 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114192:114352 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114192:114352 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114192:114352 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114192:114352 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574288:2574447 [6] NCCL INFO comm 0x8815b80 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-73:2574282:2574445 [0] NCCL INFO comm 0xa259700 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-73:2574284:2574444 [2] NCCL INFO comm 0xa03fb30 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-73:2574286:2574442 [4] NCCL INFO comm 0x8cc3430 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-73:2574283:2574443 [1] NCCL INFO comm 0x8786fe0 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-73:2574287:2574448 [5] NCCL INFO comm 0x9bf8090 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-73:2574285:2574441 [3] NCCL INFO comm 0x8691790 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-73:2574289:2574446 [7] NCCL INFO comm 0x9ab1f90 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-142:114195:114355 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114195:114355 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114195:114355 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114195:114355 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114197:114357 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114197:114357 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114197:114357 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114197:114357 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114194:114353 [2] NCCL INFO comm 0x94694e0 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-142:114192:114352 [0] NCCL INFO comm 0x99232a0 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-142:114196:114358 [4] NCCL INFO comm 0x8953d90 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-142:114193:114354 [1] NCCL INFO comm 0x9968620 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-142:114198:114351 [6] NCCL INFO comm 0x9e6bdf0 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-142:114197:114357 [5] NCCL INFO comm 0xa2e15f0 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-142:114199:114356 [7] NCCL INFO comm 0x94ed670 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-142:114195:114355 [3] NCCL INFO comm 0x987a8c0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-147:2713675:2713835 [2] NCCL INFO comm 0xa351520 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-147:2713679:2713837 [6] NCCL INFO comm 0x93f4e10 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-147:2713677:2713838 [4] NCCL INFO comm 0x939e880 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-147:2713673:2713836 [0] NCCL INFO comm 0xa05ea30 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-147:2713678:2713839 [5] NCCL INFO comm 0x9bab5e0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-147:2713680:2713833 [7] NCCL INFO comm 0x9523840 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-147:2713674:2713840 [1] NCCL INFO comm 0x991d9c0 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-147:2713676:2713834 [3] NCCL INFO comm 0x864c570 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb0728d2e94633fff - Init COMPLETE +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Using network Libfabric +ip-26-0-172-142:114196:114488 [4] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Using network Libfabric +ip-26-0-172-142:114195:114489 [3] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Using network Libfabric +ip-26-0-172-142:114194:114490 [2] NCCL INFO Using network Libfabric +ip-26-0-172-142:114193:114491 [1] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Using network Libfabric +ip-26-0-172-142:114197:114494 [5] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Using network Libfabric +ip-26-0-172-142:114198:114493 [6] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Using network Libfabric +ip-26-0-172-142:114192:114492 [0] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Using network Libfabric +ip-26-0-172-142:114199:114495 [7] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO comm 0x952ef10 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO comm 0x99b09a0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO comm 0xa5b2df0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO comm 0xa04fc00 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-142:114196:114488 [4] NCCL INFO comm 0x8b85b00 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-142:114195:114489 [3] NCCL INFO comm 0x9aacee0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-142:114194:114490 [2] NCCL INFO comm 0x969a730 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO comm 0x9ce8d40 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO comm 0x8a4ce80 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO comm 0x9e2ed80 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO comm 0x8ef9bc0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO comm 0x88c44a0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO comm 0x99f29e0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO comm 0xa276c20 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO comm 0x94abd90 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-142:114192:114492 [0] NCCL INFO comm 0x9b56080 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-142:114193:114491 [1] NCCL INFO comm 0x9b9ade0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO comm 0x89b8240 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-142:114197:114494 [5] NCCL INFO comm 0xa511f00 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-142:114198:114493 [6] NCCL INFO comm 0xa09e340 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-142:114199:114495 [7] NCCL INFO comm 0x971fa40 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO comm 0xa48a300 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO comm 0x9755f40 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO comm 0x9625200 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO comm 0x9ddbbb0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO comm 0x95d1ea0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO comm 0xa290d80 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO comm 0x9b4fe20 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO comm 0x93ceb80 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO comm 0xa581360 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO comm 0x887ebe0 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO comm 0xa581b30 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xde94afca25aa8779 - Init START +ip-26-0-172-142:114196:114488 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114194:114490 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114195:114489 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114192:114492 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114193:114491 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114197:114494 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114199:114495 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114198:114493 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-142:114194:114490 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-142:114194:114490 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-142:114197:114494 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114197:114494 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-142:114192:114492 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-142:114192:114492 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-142:114195:114489 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-142:114195:114489 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-142:114199:114495 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114199:114495 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-142:114193:114491 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-142:114193:114491 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-142:114196:114488 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114196:114488 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-142:114198:114493 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114198:114493 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Trees [0] 26/-1/-1->25->24 [1] 26/-1/-1->25->17 [2] -1/-1/-1->25->24 [3] 26/-1/-1->25->24 [4] 26/-1/-1->25->24 [5] 26/-1/-1->25->24 [6] 26/-1/-1->25->24 [7] 26/-1/-1->25->24 [8] 26/-1/-1->25->24 [9] 26/9/-1->25->-1 [10] -1/-1/-1->25->24 [11] 26/-1/-1->25->24 [12] 26/-1/-1->25->24 [13] 26/-1/-1->25->24 [14] 26/-1/-1->25->24 [15] 26/-1/-1->25->24 +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Trees [0] 27/-1/-1->26->25 [1] 27/-1/-1->26->25 [2] 27/-1/-1->26->18 [3] -1/-1/-1->26->25 [4] 27/-1/-1->26->25 [5] 27/-1/-1->26->25 [6] 27/-1/-1->26->25 [7] 27/-1/-1->26->25 [8] 27/-1/-1->26->25 [9] 27/-1/-1->26->25 [10] 27/10/-1->26->-1 [11] -1/-1/-1->26->25 [12] 27/-1/-1->26->25 [13] 27/-1/-1->26->25 [14] 27/-1/-1->26->25 [15] 27/-1/-1->26->25 +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Trees [0] 28/-1/-1->27->26 [1] 28/-1/-1->27->26 [2] 28/-1/-1->27->26 [3] 28/-1/-1->27->19 [4] -1/-1/-1->27->26 [5] 28/-1/-1->27->26 [6] 28/-1/-1->27->26 [7] 28/-1/-1->27->26 [8] 28/-1/-1->27->26 [9] 28/-1/-1->27->26 [10] 28/-1/-1->27->26 [11] 28/11/-1->27->-1 [12] -1/-1/-1->27->26 [13] 28/-1/-1->27->26 [14] 28/-1/-1->27->26 [15] 28/-1/-1->27->26 +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114199:114495 [7] NCCL INFO Trees [0] -1/-1/-1->15->14 [1] 8/-1/-1->15->14 [2] 8/-1/-1->15->14 [3] 8/-1/-1->15->14 [4] 8/-1/-1->15->14 [5] 8/-1/-1->15->14 [6] 8/-1/-1->15->14 [7] 8/-1/-1->15->23 [8] -1/-1/-1->15->14 [9] 8/-1/-1->15->14 [10] 8/-1/-1->15->14 [11] 8/-1/-1->15->14 [12] 8/-1/-1->15->14 [13] 8/-1/-1->15->14 [14] 8/-1/-1->15->14 [15] 8/23/7->15->31 +ip-26-0-172-142:114199:114495 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114198:114493 [6] NCCL INFO Trees [0] 15/-1/-1->14->13 [1] 15/-1/-1->14->13 [2] 15/-1/-1->14->13 [3] 15/-1/-1->14->13 [4] 15/-1/-1->14->13 [5] 15/-1/-1->14->13 [6] 15/-1/-1->14->22 [7] -1/-1/-1->14->13 [8] 15/-1/-1->14->13 [9] 15/-1/-1->14->13 [10] 15/-1/-1->14->13 [11] 15/-1/-1->14->13 [12] 15/-1/-1->14->13 [13] 15/-1/-1->14->13 [14] 15/22/6->14->30 [15] -1/-1/-1->14->13 +ip-26-0-172-142:114198:114493 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Trees [0] 29/-1/-1->28->27 [1] 29/-1/-1->28->27 [2] 29/-1/-1->28->27 [3] 29/-1/-1->28->27 [4] 29/-1/-1->28->20 [5] -1/-1/-1->28->27 [6] 29/-1/-1->28->27 [7] 29/-1/-1->28->27 [8] 29/-1/-1->28->27 [9] 29/-1/-1->28->27 [10] 29/-1/-1->28->27 [11] 29/-1/-1->28->27 [12] 29/12/-1->28->-1 [13] -1/-1/-1->28->27 [14] 29/-1/-1->28->27 [15] 29/-1/-1->28->27 +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Trees [0] 30/-1/-1->29->28 [1] 30/-1/-1->29->28 [2] 30/-1/-1->29->28 [3] 30/-1/-1->29->28 [4] 30/-1/-1->29->28 [5] 30/-1/-1->29->21 [6] -1/-1/-1->29->28 [7] 30/-1/-1->29->28 [8] 30/-1/-1->29->28 [9] 30/-1/-1->29->28 [10] 30/-1/-1->29->28 [11] 30/-1/-1->29->28 [12] 30/-1/-1->29->28 [13] 30/13/-1->29->-1 [14] -1/-1/-1->29->28 [15] 30/-1/-1->29->28 +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-142:114196:114488 [4] NCCL INFO Trees [0] 13/-1/-1->12->11 [1] 13/-1/-1->12->11 [2] 13/-1/-1->12->11 [3] 13/-1/-1->12->11 [4] 13/-1/-1->12->20 [5] -1/-1/-1->12->11 [6] 13/-1/-1->12->11 [7] 13/-1/-1->12->11 [8] 13/-1/-1->12->11 [9] 13/-1/-1->12->11 [10] 13/-1/-1->12->11 [11] 13/-1/-1->12->11 [12] 13/20/4->12->28 [13] -1/-1/-1->12->11 [14] 13/-1/-1->12->11 [15] 13/-1/-1->12->11 +ip-26-0-172-142:114196:114488 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Trees [0] 23/-1/-1->22->21 [1] 23/-1/-1->22->21 [2] 23/-1/-1->22->21 [3] 23/-1/-1->22->21 [4] 23/-1/-1->22->21 [5] 23/-1/-1->22->21 [6] 23/14/30->22->6 [7] -1/-1/-1->22->21 [8] 23/-1/-1->22->21 [9] 23/-1/-1->22->21 [10] 23/-1/-1->22->21 [11] 23/-1/-1->22->21 [12] 23/-1/-1->22->21 [13] 23/-1/-1->22->21 [14] 23/-1/-1->22->14 [15] -1/-1/-1->22->21 +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Trees [0] 22/-1/-1->21->20 [1] 22/-1/-1->21->20 [2] 22/-1/-1->21->20 [3] 22/-1/-1->21->20 [4] 22/-1/-1->21->20 [5] 22/13/29->21->5 [6] -1/-1/-1->21->20 [7] 22/-1/-1->21->20 [8] 22/-1/-1->21->20 [9] 22/-1/-1->21->20 [10] 22/-1/-1->21->20 [11] 22/-1/-1->21->20 [12] 22/-1/-1->21->20 [13] 22/-1/-1->21->13 [14] -1/-1/-1->21->20 [15] 22/-1/-1->21->20 +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Trees [0] 31/-1/-1->30->29 [1] 31/-1/-1->30->29 [2] 31/-1/-1->30->29 [3] 31/-1/-1->30->29 [4] 31/-1/-1->30->29 [5] 31/-1/-1->30->29 [6] 31/-1/-1->30->22 [7] -1/-1/-1->30->29 [8] 31/-1/-1->30->29 [9] 31/-1/-1->30->29 [10] 31/-1/-1->30->29 [11] 31/-1/-1->30->29 [12] 31/-1/-1->30->29 [13] 31/-1/-1->30->29 [14] 31/14/-1->30->-1 [15] -1/-1/-1->30->29 +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] 0/-1/-1->7->6 [2] 0/-1/-1->7->6 [3] 0/-1/-1->7->6 [4] 0/-1/-1->7->6 [5] 0/-1/-1->7->6 [6] 0/-1/-1->7->6 [7] 0/23/-1->7->-1 [8] -1/-1/-1->7->6 [9] 0/-1/-1->7->6 [10] 0/-1/-1->7->6 [11] 0/-1/-1->7->6 [12] 0/-1/-1->7->6 [13] 0/-1/-1->7->6 [14] 0/-1/-1->7->6 [15] 0/-1/-1->7->15 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-142:114193:114491 [1] NCCL INFO Trees [0] 10/-1/-1->9->8 [1] 10/-1/-1->9->17 [2] -1/-1/-1->9->8 [3] 10/-1/-1->9->8 [4] 10/-1/-1->9->8 [5] 10/-1/-1->9->8 [6] 10/-1/-1->9->8 [7] 10/-1/-1->9->8 [8] 10/-1/-1->9->8 [9] 10/17/1->9->25 [10] -1/-1/-1->9->8 [11] 10/-1/-1->9->8 [12] 10/-1/-1->9->8 [13] 10/-1/-1->9->8 [14] 10/-1/-1->9->8 [15] 10/-1/-1->9->8 +ip-26-0-172-142:114192:114492 [0] NCCL INFO Trees [0] 9/-1/-1->8->16 [1] -1/-1/-1->8->15 [2] 9/-1/-1->8->15 [3] 9/-1/-1->8->15 [4] 9/-1/-1->8->15 [5] 9/-1/-1->8->15 [6] 9/-1/-1->8->15 [7] 9/-1/-1->8->15 [8] 9/16/0->8->24 [9] -1/-1/-1->8->15 [10] 9/-1/-1->8->15 [11] 9/-1/-1->8->15 [12] 9/-1/-1->8->15 [13] 9/-1/-1->8->15 [14] 9/-1/-1->8->15 [15] 9/-1/-1->8->15 +ip-26-0-172-142:114193:114491 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114192:114492 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Trees [0] -1/-1/-1->31->30 [1] 24/-1/-1->31->30 [2] 24/-1/-1->31->30 [3] 24/-1/-1->31->30 [4] 24/-1/-1->31->30 [5] 24/-1/-1->31->30 [6] 24/-1/-1->31->30 [7] 24/-1/-1->31->23 [8] -1/-1/-1->31->30 [9] 24/-1/-1->31->30 [10] 24/-1/-1->31->30 [11] 24/-1/-1->31->30 [12] 24/-1/-1->31->30 [13] 24/-1/-1->31->30 [14] 24/-1/-1->31->30 [15] 24/15/-1->31->-1 +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-142:114195:114489 [3] NCCL INFO Trees [0] 12/-1/-1->11->10 [1] 12/-1/-1->11->10 [2] 12/-1/-1->11->10 [3] 12/-1/-1->11->19 [4] -1/-1/-1->11->10 [5] 12/-1/-1->11->10 [6] 12/-1/-1->11->10 [7] 12/-1/-1->11->10 [8] 12/-1/-1->11->10 [9] 12/-1/-1->11->10 [10] 12/-1/-1->11->10 [11] 12/19/3->11->27 [12] -1/-1/-1->11->10 [13] 12/-1/-1->11->10 [14] 12/-1/-1->11->10 [15] 12/-1/-1->11->10 +ip-26-0-172-142:114195:114489 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Trees [0] -1/-1/-1->23->22 [1] 16/-1/-1->23->22 [2] 16/-1/-1->23->22 [3] 16/-1/-1->23->22 [4] 16/-1/-1->23->22 [5] 16/-1/-1->23->22 [6] 16/-1/-1->23->22 [7] 16/15/31->23->7 [8] -1/-1/-1->23->22 [9] 16/-1/-1->23->22 [10] 16/-1/-1->23->22 [11] 16/-1/-1->23->22 [12] 16/-1/-1->23->22 [13] 16/-1/-1->23->22 [14] 16/-1/-1->23->22 [15] 16/-1/-1->23->15 +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Trees [0] 25/-1/-1->24->16 [1] -1/-1/-1->24->31 [2] 25/-1/-1->24->31 [3] 25/-1/-1->24->31 [4] 25/-1/-1->24->31 [5] 25/-1/-1->24->31 [6] 25/-1/-1->24->31 [7] 25/-1/-1->24->31 [8] 25/8/-1->24->-1 [9] -1/-1/-1->24->31 [10] 25/-1/-1->24->31 [11] 25/-1/-1->24->31 [12] 25/-1/-1->24->31 [13] 25/-1/-1->24->31 [14] 25/-1/-1->24->31 [15] 25/-1/-1->24->31 +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/21/-1->5->-1 [6] -1/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->13 [14] -1/-1/-1->5->4 [15] 6/-1/-1->5->4 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NVLS Head 0: 0 8 16 24 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-142:114197:114494 [5] NCCL INFO Trees [0] 14/-1/-1->13->12 [1] 14/-1/-1->13->12 [2] 14/-1/-1->13->12 [3] 14/-1/-1->13->12 [4] 14/-1/-1->13->12 [5] 14/-1/-1->13->21 [6] -1/-1/-1->13->12 [7] 14/-1/-1->13->12 [8] 14/-1/-1->13->12 [9] 14/-1/-1->13->12 [10] 14/-1/-1->13->12 [11] 14/-1/-1->13->12 [12] 14/-1/-1->13->12 [13] 14/21/5->13->29 [14] -1/-1/-1->13->12 [15] 14/-1/-1->13->12 +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Trees [0] 18/-1/-1->17->16 [1] 18/9/25->17->1 [2] -1/-1/-1->17->16 [3] 18/-1/-1->17->16 [4] 18/-1/-1->17->16 [5] 18/-1/-1->17->16 [6] 18/-1/-1->17->16 [7] 18/-1/-1->17->16 [8] 18/-1/-1->17->16 [9] 18/-1/-1->17->9 [10] -1/-1/-1->17->16 [11] 18/-1/-1->17->16 [12] 18/-1/-1->17->16 [13] 18/-1/-1->17->16 [14] 18/-1/-1->17->16 [15] 18/-1/-1->17->16 +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Trees [0] 20/-1/-1->19->18 [1] 20/-1/-1->19->18 [2] 20/-1/-1->19->18 [3] 20/11/27->19->3 [4] -1/-1/-1->19->18 [5] 20/-1/-1->19->18 [6] 20/-1/-1->19->18 [7] 20/-1/-1->19->18 [8] 20/-1/-1->19->18 [9] 20/-1/-1->19->18 [10] 20/-1/-1->19->18 [11] 20/-1/-1->19->11 [12] -1/-1/-1->19->18 [13] 20/-1/-1->19->18 [14] 20/-1/-1->19->18 [15] 20/-1/-1->19->18 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NVLS Head 1: 1 9 17 25 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NVLS Head 2: 2 10 18 26 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-142:114197:114494 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114194:114490 [2] NCCL INFO Trees [0] 11/-1/-1->10->9 [1] 11/-1/-1->10->9 [2] 11/-1/-1->10->18 [3] -1/-1/-1->10->9 [4] 11/-1/-1->10->9 [5] 11/-1/-1->10->9 [6] 11/-1/-1->10->9 [7] 11/-1/-1->10->9 [8] 11/-1/-1->10->9 [9] 11/-1/-1->10->9 [10] 11/18/2->10->26 [11] -1/-1/-1->10->9 [12] 11/-1/-1->10->9 [13] 11/-1/-1->10->9 [14] 11/-1/-1->10->9 [15] 11/-1/-1->10->9 +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Trees [0] 19/-1/-1->18->17 [1] 19/-1/-1->18->17 [2] 19/10/26->18->2 [3] -1/-1/-1->18->17 [4] 19/-1/-1->18->17 [5] 19/-1/-1->18->17 [6] 19/-1/-1->18->17 [7] 19/-1/-1->18->17 [8] 19/-1/-1->18->17 [9] 19/-1/-1->18->17 [10] 19/-1/-1->18->10 [11] -1/-1/-1->18->17 [12] 19/-1/-1->18->17 [13] 19/-1/-1->18->17 [14] 19/-1/-1->18->17 [15] 19/-1/-1->18->17 +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Trees [0] 17/8/24->16->0 [1] -1/-1/-1->16->23 [2] 17/-1/-1->16->23 [3] 17/-1/-1->16->23 [4] 17/-1/-1->16->23 [5] 17/-1/-1->16->23 [6] 17/-1/-1->16->23 [7] 17/-1/-1->16->23 [8] 17/-1/-1->16->8 [9] -1/-1/-1->16->23 [10] 17/-1/-1->16->23 [11] 17/-1/-1->16->23 [12] 17/-1/-1->16->23 [13] 17/-1/-1->16->23 [14] 17/-1/-1->16->23 [15] 17/-1/-1->16->23 +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NVLS Head 3: 3 11 19 27 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NVLS Head 4: 4 12 20 28 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/18/-1->2->-1 [3] -1/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->10 [11] -1/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-142:114194:114490 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Trees [0] 21/-1/-1->20->19 [1] 21/-1/-1->20->19 [2] 21/-1/-1->20->19 [3] 21/-1/-1->20->19 [4] 21/12/28->20->4 [5] -1/-1/-1->20->19 [6] 21/-1/-1->20->19 [7] 21/-1/-1->20->19 [8] 21/-1/-1->20->19 [9] 21/-1/-1->20->19 [10] 21/-1/-1->20->19 [11] 21/-1/-1->20->19 [12] 21/-1/-1->20->12 [13] -1/-1/-1->20->19 [14] 21/-1/-1->20->19 [15] 21/-1/-1->20->19 +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NVLS Head 5: 5 13 21 29 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 00/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/22/-1->6->-1 [7] -1/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->14 [15] -1/-1/-1->6->5 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NVLS Head 6: 6 14 22 30 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 01/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NVLS Head 7: 7 15 23 31 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 02/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/20/-1->4->-1 [5] -1/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->12 [13] -1/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 03/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 04/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 05/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/19/-1->3->-1 [4] -1/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->11 [12] -1/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 06/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/17/-1->1->-1 [2] -1/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->9 [10] -1/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 07/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 08/16 : 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 09/16 : 0 7 6 5 4 3 2 9 8 15 14 13 12 11 10 17 16 23 22 21 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 10/16 : 0 7 6 5 4 3 10 9 8 15 14 13 12 11 18 17 16 23 22 21 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 11/16 : 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 12/16 : 0 7 6 5 12 11 10 9 8 15 14 13 20 19 18 17 16 23 22 21 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 13/16 : 0 4 7 6 13 11 10 9 8 12 15 14 21 19 18 17 16 20 23 22 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 14/16 : 0 5 4 7 14 11 10 9 8 13 12 15 22 19 18 17 16 21 20 23 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 15/16 : 0 6 5 4 15 11 10 9 8 14 13 12 23 19 18 17 16 22 21 20 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Trees [0] 1/16/-1->0->-1 [1] -1/-1/-1->0->7 [2] 1/-1/-1->0->7 [3] 1/-1/-1->0->7 [4] 1/-1/-1->0->7 [5] 1/-1/-1->0->7 [6] 1/-1/-1->0->7 [7] 1/-1/-1->0->7 [8] 1/-1/-1->0->8 [9] -1/-1/-1->0->7 [10] 1/-1/-1->0->7 [11] 1/-1/-1->0->7 [12] 1/-1/-1->0->7 [13] 1/-1/-1->0->7 [14] 1/-1/-1->0->7 [15] 1/-1/-1->0->7 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 00/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 00/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 00/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 00/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 00/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 00/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 00/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 00/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 08/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 00/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 00/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 08/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 00/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 00/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 08/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 08/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 08/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 08/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 08/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 08/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 08/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 05/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 08/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 08/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 06/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 13/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 14/0 : 12[4] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 05/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 08/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 06/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 13/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 05/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 14/0 : 4[4] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 00/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 00/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 00/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 08/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 06/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 00/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 08/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 13/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 14/0 : 20[4] -> 23[7] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 08/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 08/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 00/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 00/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 08/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 08/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 05/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 06/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 13/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 14/0 : 28[4] -> 31[7] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 00/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 08/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [send] via NET/Libfabric/0(8)/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 00/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 08/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 00/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 08/0 : 31[7] -> 0[0] [send] via NET/Libfabric/0(24)/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 00/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 08/0 : 7[7] -> 8[0] [send] via NET/Libfabric/0(0)/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 05/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 05/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 13/0 : 8[0] -> 12[4] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 13/0 : 0[0] -> 4[4] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [send] via NET/Libfabric/2(10)/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 00/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 08/0 : 15[7] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 00/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 00/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 08/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 08/0 : 23[7] -> 24[0] [send] via NET/Libfabric/0(16)/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 05/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 05/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 13/0 : 24[0] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 13/0 : 16[0] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 06/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 14/0 : 24[0] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 07/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 15/0 : 24[0] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 01/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 06/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 06/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 02/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 14/0 : 16[0] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 03/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 14/0 : 8[0] -> 13[5] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 04/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 07/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 07/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 09/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 10/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 15/0 : 16[0] -> 22[6] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 15/0 : 8[0] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 11/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 12/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 02/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 01/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 10/0 : 3[3] -> 10[2] [send] via NET/Libfabric/2(2)/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 01/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [send] via NET/Libfabric/6(6)/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 02/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 10/0 : 11[3] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 02/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 03/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [send] via NET/Libfabric/2(18)/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 04/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [send] via NET/Libfabric/6(14)/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 02/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 09/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 03/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 10/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 06/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 04/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 09/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 14/0 : 0[0] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 10/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 11/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 07/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 12/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 11/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 15/0 : 0[0] -> 6[6] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 12/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 01/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 02/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 10/0 : 19[3] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 02/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 03/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 02/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 10/0 : 27[3] -> 2[2] [send] via NET/Libfabric/2(26)/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 04/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 09/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 10/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 11/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 12/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 01/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [send] via NET/Libfabric/1(9)/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [send] via NET/Libfabric/1(25)/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 01/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 09/0 : 10[2] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 01/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 09/0 : 18[2] -> 25[1] [send] via NET/Libfabric/1(17)/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 03/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 11/0 : 12[4] -> 19[3] [send] via NET/Libfabric/3(11)/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [send] via NET/Libfabric/4(12)/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 01/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 09/0 : 26[2] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 09/0 : 2[2] -> 9[1] [send] via NET/Libfabric/1(1)/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 03/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [send] via NET/Libfabric/6(30)/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 11/0 : 28[4] -> 3[3] [send] via NET/Libfabric/3(27)/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [send] via NET/Libfabric/4(28)/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 02/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [send] via NET/Libfabric/5(29)/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 03/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 04/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 01/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 06/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 01/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 14/0 : 7[7] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 05/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 02/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 02/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [send] via NET/Libfabric/5(13)/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 02/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 06/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 03/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 03/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 03/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 04/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 07/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 04/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 04/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 12/0 : 13[5] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 04/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 10/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 03/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 05/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 05/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 11/0 : 20[4] -> 27[3] [send] via NET/Libfabric/3(19)/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 05/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 06/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 06/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 11/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 14/0 : 15[7] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 05/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 06/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 06/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 13/0 : 14[6] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 14/0 : 23[7] -> 30[6] [send] via NET/Libfabric/6(22)/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 06/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 12/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 05/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 04/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 13/0 : 22[6] -> 29[5] [send] via NET/Libfabric/5(21)/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 07/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 07/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 12/0 : 21[5] -> 28[4] [send] via NET/Libfabric/4(20)/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 13/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 06/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 06/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 05/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 09/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 07/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 10/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 14/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 09/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 14/0 : 14[6] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 04/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 05/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 15/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 12/0 : 29[5] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 13/0 : 30[6] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 10/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 06/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 14/0 : 22[6] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 03/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 10/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 01/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 11/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 13/0 : 21[5] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 04/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 11/0 : 4[4] -> 11[3] [send] via NET/Libfabric/3(3)/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 14/0 : 31[7] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 12/0 : 5[5] -> 12[4] [send] via NET/Libfabric/4(4)/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 11/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 02/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 05/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 02/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 05/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 06/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 12/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 13/0 : 6[6] -> 13[5] [send] via NET/Libfabric/5(5)/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 05/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 03/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 12/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 11/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 06/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 05/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 03/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 13/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 14/0 : 30[6] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 12/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 13/0 : 29[5] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 13/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 13/0 : 5[5] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 14/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 13/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 04/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 04/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 14/0 : 6[6] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 15/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 14/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 14/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 05/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 05/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 13/0 : 13[5] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 15/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 15/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 06/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 06/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 07/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 09/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 07/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 10/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 10/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 11/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 11/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 12/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 13/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 12/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 14/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 15/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 13/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 14/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 15/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 07/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [send] via NET/Libfabric/7(7)/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 07/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 15/0 : 28[4] -> 7[7] [send] via NET/Libfabric/7(31)/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 07/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 15/0 : 4[4] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 07/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [send] via NET/Libfabric/7(15)/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 07/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 07/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 15/0 : 20[4] -> 31[7] [send] via NET/Libfabric/7(23)/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 15/0 : 12[4] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 07/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 07/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 15/0 : 7[7] -> 3[3] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 15/0 : 15[7] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 15/0 : 23[7] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 15/0 : 31[7] -> 27[3] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 01/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 01/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 01/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 02/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 03/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 02/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 04/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 02/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 03/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 05/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 04/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 09/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 05/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 10/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 03/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 09/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 11/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 04/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 10/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 05/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 11/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 12/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 01/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 12/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 01/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 09/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 13/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 02/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 13/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 02/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 10/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 04/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 04/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 11/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 09/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 09/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 01/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 10/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 12/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 02/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 10/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 12/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 01/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 13/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 04/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 12/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 09/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 03/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 01/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 10/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 04/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 01/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 12/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 02/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 01/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 02/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 03/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 04/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 03/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 07/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 03/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 09/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 04/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 01/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 04/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 10/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 03/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 11/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 05/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 07/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 05/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 12/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 04/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 01/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 06/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 15/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 06/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 05/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 02/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 07/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 07/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 06/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 09/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 03/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 11/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 07/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 01/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 09/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 01/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 04/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 12/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 09/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 02/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 10/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 02/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 07/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 13/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 11/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 09/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 03/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 03/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 11/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 09/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 14/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 12/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 06/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 06/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 12/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 10/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 15/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 13/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 07/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 11/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 07/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 15/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 14/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 09/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 09/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 01/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 10/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 12/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 11/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 15/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 15/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 11/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 14/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 15/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 02/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 12/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 03/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 13/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 10/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 06/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 14/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 15/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 07/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 09/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 11/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 10/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 14/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 11/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 15/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 14/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 15/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Connected all rings +ip-26-0-172-142:114193:114491 [1] NCCL INFO Connected all rings +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Connected all rings +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 01/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 03/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 04/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 05/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 06/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 07/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 09/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 11/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 12/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 13/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 14/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 15/0 : 9[1] -> 10[2] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 02/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 03/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 04/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 05/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 06/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 07/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 10/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 11/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 12/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 13/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 14/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 15/0 : 24[0] -> 25[1] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Connected all rings +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 02/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 01/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 02/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 03/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Connected all rings +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 04/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 05/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 06/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 01/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 03/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 04/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 05/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 09/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Connected all rings +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 10/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 06/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 07/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Connected all rings +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 11/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Connected all rings +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 09/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 12/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Connected all rings +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 11/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 13/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 12/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 14/0 : 14[6] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 13/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Connected all rings +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 14/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 15/0 : 25[1] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 05/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 06/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 07/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 01/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 13/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 01/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 02/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 14/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 02/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 04/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 01/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Connected all rings +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 15/0 : 24[0] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 03/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 05/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 02/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 02/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 04/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 06/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 03/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 03/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 06/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 07/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 04/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 07/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 05/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 09/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Connected all rings +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 09/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 07/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 04/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 05/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 06/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 07/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 10/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 10/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 10/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 09/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 12/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 11/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 01/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 10/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 13/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 12/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 11/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 12/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 13/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 11/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 14/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 14/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 15/0 : 8[0] -> 9[1] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 12/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 15/0 : 26[2] -> 27[3] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 01/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 02/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 05/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 14/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 13/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 03/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 06/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 15/0 : 28[4] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 15/0 : 29[5] -> 30[6] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 04/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 07/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 05/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 13/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 01/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 02/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 07/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 14/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 03/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 04/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 15/0 : 8[0] -> 15[7] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 09/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 05/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 06/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 10/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 09/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 11/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 10/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 12/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 01/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 11/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 12/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 13/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Connected all rings +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 13/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 14/0 : 30[6] -> 31[7] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 02/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Connected all rings +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 03/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 04/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 15/0 : 13[5] -> 14[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 06/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 07/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 09/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 10/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 11/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 12/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 14/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 15/0 : 12[4] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 03/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 01/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 02/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 01/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 04/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Connected all rings +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 02/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 05/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 03/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 04/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 06/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 05/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 07/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 05/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 05/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 06/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 09/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 07/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 06/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 10/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 06/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 07/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 13/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 07/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 14/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Connected all rings +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 09/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 12/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 15/0 : 0[0] -> 7[7] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 10/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 13/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 11/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 14/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 13/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 15/0 : 10[2] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 14/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 10/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 09/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 15/0 : 11[3] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Connected all rings +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 10/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 13/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 11/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 12/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 11/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 14/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 12/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Connected all rings +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Connected all rings +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 15/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 08/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 13/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 14/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 15/0 : 16[0] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Connected all rings +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Connected all rings +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 01/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 02/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 03/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 05/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 06/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 07/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 09/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 10/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 11/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 13/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 14/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 15/0 : 27[3] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 01/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 01/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 02/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 03/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 04/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 04/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 05/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 05/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 06/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 07/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 09/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 01/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 10/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 02/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 12/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 01/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 03/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 13/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 02/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 06/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 04/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 14/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 03/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 07/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 05/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 15/0 : 18[2] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 04/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 09/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 06/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 05/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 11/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 09/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 07/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 12/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 10/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 09/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 13/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 11/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 10/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 14/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 12/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 02/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 11/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Connected all rings +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 15/0 : 17[1] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 13/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 03/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 12/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 14/0 : 22[6] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 05/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 01/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 09/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 13/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 05/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 01/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 01/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 01/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 06/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 15/0 : 21[5] -> 22[6] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 06/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 09/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 09/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 01/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 01/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 07/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 07/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 06/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 14/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 06/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 13/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 09/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 06/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 14/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 06/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 14/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 00/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 14/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 09/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 06/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 10/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 08/0 : 25[1] -> 24[0] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 06/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 09/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 14/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 14/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 15/0 : 16[0] -> 23[7] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 11/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 00/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 05/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 08/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 13/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 13/0 : 30[6] -> 29[5] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 14/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 00/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 07/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 08/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 00/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 15/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 00/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 07/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 15/0 : 19[3] -> 20[4] via P2P/IPC +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 08/0 : 9[1] -> 8[0] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 00/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 00/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 07/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 00/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 07/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 00/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 00/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 08/0 : 17[1] -> 16[0] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 05/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 02/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 05/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 10/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 02/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 08/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 13/0 : 14[6] -> 13[5] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 01/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 08/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 02/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 02/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 10/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 10/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 02/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 13/0 : 22[6] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 02/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 02/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 08/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 10/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 03/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 08/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 10/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 04/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 00/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 00/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 08/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 08/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 06/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 07/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 15/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 15/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 09/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 10/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 00/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 11/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 12/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 01/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 14/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 08/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 15/0 : 20[4] -> 21[5] via P2P/IPC +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 09/0 : 10[2] -> 9[1] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 03/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 05/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 11/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 04/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 13/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 03/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 12/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 05/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 04/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 07/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 15/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 15/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 03/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 05/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 12/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 03/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 04/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 05/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 12/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 01/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 04/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 04/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 00/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 04/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 13/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 13/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 12/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 02/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 12/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 01/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 05/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 05/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 07/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 03/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 13/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 13/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 01/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 01/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 11/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 11/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 08/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 04/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 03/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 05/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 00/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 02/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 03/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 11/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 11/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 02/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 06/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 09/0 : 26[2] -> 25[1] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 07/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 01/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 03/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 03/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 09/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 04/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 08/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 00/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 10/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 04/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 00/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 03/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 11/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 00/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 05/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 03/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 00/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 05/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 12/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 05/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 09/0 : 18[2] -> 17[1] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 04/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 06/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 05/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 00/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 03/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 06/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 13/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 04/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 08/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 05/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 07/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 06/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 06/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 14/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 07/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 08/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 06/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 12/0 : 29[5] -> 28[4] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 09/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 07/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 07/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 15/0 : 15[7] -> 8[0] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 12/0 : 13[5] -> 12[4] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 07/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 09/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 10/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 00/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 08/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 00/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 08/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 08/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 11/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 10/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 11/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 12/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 06/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 11/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 11/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 00/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 13/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 13/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 04/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 02/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 14/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 14/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 08/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 11/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 15/0 : 28[4] -> 27[3] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 15/0 : 31[7] -> 24[0] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 13/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 10/0 : 27[3] -> 26[2] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 00/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 13/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 08/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 12/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 00/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 14/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 00/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 13/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 02/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 15/0 : 20[4] -> 19[3] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 14/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 08/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 08/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 15/0 : 23[7] -> 16[0] via P2P/IPC +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 12/0 : 21[5] -> 20[4] via P2P/IPC +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 10/0 : 19[3] -> 18[2] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 14/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 14/0 : 15[7] -> 14[6] via P2P/IPC +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 15/0 : 12[4] -> 11[3] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 06/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 02/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 00/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 08/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 08/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 06/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 10/0 : 11[3] -> 10[2] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 08/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 14/0 : 31[7] -> 30[6] via P2P/IPC +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 14/0 : 23[7] -> 22[6] via P2P/IPC +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Connected all trees +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NVLS comm 0x99b09a0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Connected all trees +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO NVLS comm 0x8ef9bc0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Connected all trees +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NVLS comm 0xa04fc00 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Connected all trees +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NVLS comm 0x952ef10 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Connected all trees +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NVLS comm 0x93ceb80 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Connected all trees +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NVLS comm 0xa5b2df0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Connected all trees +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Connected all trees +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NVLS comm 0x99f29e0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NVLS comm 0x94abd90 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Connected all trees +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO NVLS comm 0xa276c20 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Connected all trees +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NVLS comm 0xa581360 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Connected all trees +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO NVLS comm 0x88c44a0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Connected all trees +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO NVLS comm 0xa48a300 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Connected all trees +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO NVLS comm 0x9e2ed80 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114196:114488 [4] NCCL INFO Connected all trees +ip-26-0-172-142:114196:114488 [4] NCCL INFO NVLS comm 0x8b85b00 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Connected all trees +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO NVLS comm 0x8a4ce80 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Connected all trees +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO NVLS comm 0x9ce8d40 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Connected all trees +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO NVLS comm 0x89b8240 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114193:114491 [1] NCCL INFO Connected all trees +ip-26-0-172-142:114193:114491 [1] NCCL INFO NVLS comm 0x9b9ade0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114199:114495 [7] NCCL INFO Connected all trees +ip-26-0-172-142:114199:114495 [7] NCCL INFO NVLS comm 0x971fa40 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114197:114494 [5] NCCL INFO Connected all trees +ip-26-0-172-142:114197:114494 [5] NCCL INFO NVLS comm 0xa511f00 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Connected all trees +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO NVLS comm 0x9ddbbb0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114198:114493 [6] NCCL INFO Connected all trees +ip-26-0-172-142:114198:114493 [6] NCCL INFO NVLS comm 0xa09e340 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114194:114490 [2] NCCL INFO Connected all trees +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Connected all trees +ip-26-0-172-142:114194:114490 [2] NCCL INFO NVLS comm 0x969a730 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO NVLS comm 0x9625200 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Connected all trees +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO NVLS comm 0x9755f40 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Connected all trees +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO NVLS comm 0x95d1ea0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114195:114489 [3] NCCL INFO Connected all trees +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Connected all trees +ip-26-0-172-142:114195:114489 [3] NCCL INFO NVLS comm 0x9aacee0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114192:114492 [0] NCCL INFO Connected all trees +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Connected all trees +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO NVLS comm 0xa581b30 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114192:114492 [0] NCCL INFO NVLS comm 0x9b56080 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO NVLS comm 0x9b4fe20 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Connected all trees +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO NVLS comm 0x887ebe0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Connected all trees +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO NVLS comm 0xa290d80 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 02/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 00/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 04/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 00/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 04/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 06/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 02/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 08/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 06/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 04/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 00/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 08/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 10/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 00/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 08/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 02/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 10/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 12/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 00/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 02/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 10/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 04/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 00/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 12/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 14/0 : 16[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 04/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 12/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 06/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 00/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 14/0 : 18[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 02/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 02/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 14/0 : 22[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 06/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 08/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 02/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 04/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 04/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 08/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 10/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 06/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 06/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 06/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 10/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 12/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 08/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 08/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 08/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 01/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 12/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 14/0 : 19[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 10/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 10/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 03/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 10/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 14/0 : 21[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 12/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 01/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 01/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 12/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 12/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 14/0 : 20[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 05/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 14/0 : 23[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 03/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 03/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 14/0 : 17[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 01/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 07/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 01/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 01/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 05/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 05/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 01/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 03/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 09/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 03/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 03/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 01/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 07/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 07/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 05/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 03/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 11/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 05/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 05/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 03/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 09/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 09/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 13/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 05/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 07/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 11/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 07/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 05/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 07/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 11/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 15/0 : 6[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 09/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 07/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 13/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 11/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 07/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 09/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 00/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 11/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 09/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 13/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 15/0 : 0[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 13/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 09/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 11/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 01/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 13/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 11/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 15/0 : 2[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 01/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 15/0 : 1[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 13/0 : 7[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 13/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 02/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 15/0 : 4[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 15/0 : 5[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 02/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 00/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 00/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 00/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 15/0 : 3[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 00/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 00/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 03/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 01/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 03/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 02/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 01/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 00/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 01/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 01/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 04/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 04/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 03/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 02/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 03/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 02/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 01/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 02/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 05/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 04/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 03/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 05/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 04/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 03/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 02/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 03/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 07/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 05/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 04/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 05/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 06/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 04/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 05/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 04/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 08/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 06/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 05/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 06/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 06/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 07/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 06/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 05/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 09/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 07/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 06/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 07/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 09/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 07/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 07/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 10/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 06/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 08/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 08/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 08/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 10/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 08/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 08/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 11/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 07/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 09/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 09/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 10/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 11/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 09/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 09/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 12/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 10/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 08/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 10/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 11/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 10/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 12/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 11/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 13/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 11/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 09/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 12/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 11/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 11/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 13/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 15/0 : 14[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 12/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 12/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 10/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 12/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 13/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 13/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 14/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 13/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 12/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 13/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 14/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 14/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 14/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 15/0 : 8[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 14/0 : 15[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 13/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 15/0 : 13[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 14/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 15/0 : 9[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 15/0 : 12[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 14/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 15/0 : 10[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 15/0 : 11[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 01/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 03/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 05/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 01/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 01/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 07/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 01/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 01/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 01/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 03/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 03/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 01/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 09/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 03/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 01/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 03/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 03/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 05/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 05/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 11/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 05/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 03/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 03/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 05/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 05/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 07/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 13/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 07/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 07/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 05/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 05/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 07/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 07/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 09/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 15/0 : 30[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 09/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 09/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 07/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 07/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 11/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 09/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 11/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 01/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 11/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 11/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 00/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 09/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 09/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 11/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 13/0 : 31[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 13/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 03/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 15/0 : 29[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 13/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 13/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 02/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 11/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 13/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 01/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 00/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 02/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 02/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 04/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 15/0 : 25[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 05/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 01/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 15/0 : 27[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 15/0 : 24[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 13/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 15/0 : 28[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 02/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 03/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 04/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 07/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 00/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 01/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 00/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 00/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 00/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 03/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 04/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 04/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 08/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 00/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 01/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 01/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 15/0 : 26[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 01/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 05/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 06/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 09/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 05/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 03/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 00/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 00/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 02/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 04/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 02/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 08/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 02/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 08/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 00/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 03/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 03/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 07/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 03/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 01/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 06/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 10/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 04/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 06/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 00/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 00/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 06/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 00/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 11/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 07/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 02/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 05/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 02/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 04/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 10/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 08/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 10/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 00/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 04/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 05/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 09/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 05/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 12/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 02/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 02/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 06/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 08/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 02/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 04/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 05/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 13/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 03/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 09/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 07/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 08/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 02/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 06/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 06/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 12/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 10/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 14/0 : 22[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 08/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 04/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 12/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 11/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 02/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 10/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 04/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 10/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 06/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 06/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 08/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 14/0 : 6[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 04/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 04/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 12/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 08/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 14/0 : 0[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 06/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 10/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 07/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 07/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 07/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 08/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 11/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 15/0 : 14[6] -> 30[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 11/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 05/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 04/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 06/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 12/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 08/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 10/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 12/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 06/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 14/0 : 16[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 06/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 10/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 13/0 : 15[7] -> 31[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 09/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 09/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 06/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 08/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 12/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 08/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 10/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 08/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 14/0 : 4[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 09/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 10/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 14/0 : 2[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 15/0 : 13[5] -> 29[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 13/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 07/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 12/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 11/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 12/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 08/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 10/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 14/0 : 18[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 08/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 10/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 12/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 14/0 : 5[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 11/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 13/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 12/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 09/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 15/0 : 9[1] -> 25[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 12/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 13/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 10/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 10/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 00/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 14/0 : 7[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 14/0 : 1[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 14/0 : 20[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 10/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 12/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 14/0 : 17[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 12/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 02/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 12/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 13/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 15/0 : 11[3] -> 27[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 11/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 15/0 : 12[4] -> 28[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 15/0 : 8[0] -> 24[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 13/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 14/0 : 21[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 14/0 : 19[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 14/0 : 23[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 04/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 15/0 : 10[2] -> 26[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 08/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 00/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 00/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 02/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 10/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 00/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 00/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 02/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 02/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 04/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 00/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 12/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 00/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 02/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 04/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 02/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 14/0 : 30[6] -> 22[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 06/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 06/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 04/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 02/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 08/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 06/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 08/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 04/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 00/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 06/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 04/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 10/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 04/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 01/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 06/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 10/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 08/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 01/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 12/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 10/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 08/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 12/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 06/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Channel 14/0 : 24[0] -> 16[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 02/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 06/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 12/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 08/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 02/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Channel 14/0 : 26[2] -> 18[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 10/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 00/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 14/0 : 28[4] -> 20[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 08/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 03/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 08/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 12/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 03/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 10/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 01/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 00/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 10/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 04/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 12/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 14/0 : 25[1] -> 17[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 04/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 10/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 03/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 01/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 05/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 12/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 14/0 : 29[5] -> 21[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 00/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 05/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 04/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 12/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 02/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 06/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 14/0 : 31[7] -> 23[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 00/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 02/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 07/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 01/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 05/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 14/0 : 27[3] -> 19[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 07/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 03/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 00/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 01/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 03/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 08/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 06/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 00/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 09/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 03/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 05/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 02/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 04/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 01/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 07/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 09/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 01/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 10/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 06/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 03/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 05/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 08/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 02/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 02/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 11/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 10/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 07/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 04/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 06/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 09/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 03/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 04/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 12/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 11/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 06/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 08/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 07/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 11/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 04/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 05/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 13/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 12/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 07/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 09/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 12/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 08/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 05/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 06/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 14/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 13/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 08/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114495 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 13/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 10/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 10/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 07/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Channel 15/0 : 16[0] -> 8[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 06/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Channel 15/0 : 22[6] -> 14[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 09/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 14/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114198:114493 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 08/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 11/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 11/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 08/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114197:114494 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 10/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Channel 15/0 : 18[2] -> 10[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 09/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 13/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 12/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 09/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 11/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 10/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 14/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 10/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 13/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 12/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 12/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Channel 15/0 : 20[4] -> 12[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 11/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114489 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114488 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114193:114491 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 14/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 13/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 14/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 12/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114492 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Channel 15/0 : 17[1] -> 9[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 14/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Channel 15/0 : 21[5] -> 13[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 13/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Channel 15/0 : 19[3] -> 11[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Channel 14/0 : 23[7] -> 15[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114194:114490 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 01/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 12/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 05/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 01/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 03/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 14/0 : 3[3] -> 19[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 07/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 03/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 01/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 01/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 09/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 05/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 05/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 03/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 05/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 07/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 07/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 09/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 09/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 11/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 11/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 11/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 13/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 13/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 13/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Channel 15/0 : 14[6] -> 6[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Channel 15/0 : 12[4] -> 4[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Channel 15/0 : 10[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 07/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 09/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 03/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 11/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 05/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 13/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 01/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 01/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 07/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Channel 15/0 : 8[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 01/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 03/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 03/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 11/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 03/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 13/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 05/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 05/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 05/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Channel 15/0 : 9[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 07/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 07/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 07/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 09/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 09/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 09/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 13/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 11/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 11/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Channel 15/0 : 13[5] -> 5[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Channel 13/0 : 15[7] -> 7[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Channel 15/0 : 11[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888255:3888552 [7] NCCL INFO comm 0x952ef10 rank 7 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-116:3888253:3888551 [5] NCCL INFO comm 0x99b09a0 rank 5 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-116:3888251:3888555 [3] NCCL INFO comm 0x99f29e0 rank 3 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-116:3888250:3888553 [2] NCCL INFO comm 0x94abd90 rank 2 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-116:3888249:3888556 [1] NCCL INFO comm 0x93ceb80 rank 1 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-116:3888252:3888554 [4] NCCL INFO comm 0xa04fc00 rank 4 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-116:3888254:3888550 [6] NCCL INFO comm 0xa5b2df0 rank 6 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-116:3888248:3888549 [0] NCCL INFO comm 0xa581360 rank 0 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114194:114490 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114194:114490 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114194:114490 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114194:114490 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114198:114493 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114198:114493 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114198:114493 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114198:114493 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114196:114488 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114196:114488 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114196:114488 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114196:114488 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114192:114492 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114192:114492 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114192:114492 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114192:114492 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114199:114495 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114199:114495 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114199:114495 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114199:114495 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114193:114491 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114193:114491 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114193:114491 [1] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114193:114491 [1] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574289:2574587 [7] NCCL INFO comm 0x9ce8d40 rank 31 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-73:2574285:2574582 [3] NCCL INFO comm 0x88c44a0 rank 27 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-73:2574287:2574586 [5] NCCL INFO comm 0x9e2ed80 rank 29 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-73:2574283:2574583 [1] NCCL INFO comm 0x89b8240 rank 25 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-73:2574286:2574580 [4] NCCL INFO comm 0x8ef9bc0 rank 28 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-73:2574288:2574585 [6] NCCL INFO comm 0x8a4ce80 rank 30 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-73:2574282:2574581 [0] NCCL INFO comm 0xa48a300 rank 24 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-73:2574284:2574584 [2] NCCL INFO comm 0xa276c20 rank 26 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-142:114195:114489 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114195:114489 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114195:114489 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114195:114489 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114197:114494 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114197:114494 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114197:114494 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-142:114197:114494 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO threadThresholds 8/8/64 | 256/8/64 | 512 | 512 +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO 16 coll channels, 16 nvls channels, 16 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114198:114493 [6] NCCL INFO comm 0xa09e340 rank 14 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-142:114199:114495 [7] NCCL INFO comm 0x971fa40 rank 15 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-142:114195:114489 [3] NCCL INFO comm 0x9aacee0 rank 11 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-142:114193:114491 [1] NCCL INFO comm 0x9b9ade0 rank 9 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-142:114194:114490 [2] NCCL INFO comm 0x969a730 rank 10 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-142:114197:114494 [5] NCCL INFO comm 0xa511f00 rank 13 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-142:114192:114492 [0] NCCL INFO comm 0x9b56080 rank 8 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-142:114196:114488 [4] NCCL INFO comm 0x8b85b00 rank 12 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-147:2713675:2713977 [2] NCCL INFO comm 0xa581b30 rank 18 nranks 32 cudaDev 2 nvmlDev 2 busId 75000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-147:2713679:2713973 [6] NCCL INFO comm 0x9625200 rank 22 nranks 32 cudaDev 6 nvmlDev 6 busId b9000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-147:2713677:2713974 [4] NCCL INFO comm 0x95d1ea0 rank 20 nranks 32 cudaDev 4 nvmlDev 4 busId 97000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-147:2713673:2713975 [0] NCCL INFO comm 0xa290d80 rank 16 nranks 32 cudaDev 0 nvmlDev 0 busId 53000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-147:2713674:2713976 [1] NCCL INFO comm 0x9b4fe20 rank 17 nranks 32 cudaDev 1 nvmlDev 1 busId 64000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-147:2713678:2713971 [5] NCCL INFO comm 0x9ddbbb0 rank 21 nranks 32 cudaDev 5 nvmlDev 5 busId a8000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-147:2713680:2713972 [7] NCCL INFO comm 0x9755f40 rank 23 nranks 32 cudaDev 7 nvmlDev 7 busId ca000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-147:2713676:2713970 [3] NCCL INFO comm 0x887ebe0 rank 19 nranks 32 cudaDev 3 nvmlDev 3 busId 86000 commId 0xde94afca25aa8779 - Init COMPLETE +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO comm 0xa0628c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb335559d1e773894 - Init START +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO comm 0x9a056a0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb335559d1e773894 - Init START +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO comm 0x99c3660 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb335559d1e773894 - Init START +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO comm 0x94bea50 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb335559d1e773894 - Init START +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO comm 0xa5c5ab0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb335559d1e773894 - Init START +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO comm 0x9541bd0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb335559d1e773894 - Init START +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO comm 0xa594020 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb335559d1e773894 - Init START +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO comm 0x93e3960 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb335559d1e773894 - Init START +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO comm 0xa2a71c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7c40d7ed88c5d4d0 - Init START +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO comm 0x976c800 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7c40d7ed88c5d4d0 - Init START +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO comm 0x963d1d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7c40d7ed88c5d4d0 - Init START +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO comm 0x9b66f50 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7c40d7ed88c5d4d0 - Init START +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO comm 0x95e7530 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7c40d7ed88c5d4d0 - Init START +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO comm 0xa59a7e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7c40d7ed88c5d4d0 - Init START +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO comm 0x9df3bc0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7c40d7ed88c5d4d0 - Init START +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO comm 0x8895220 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7c40d7ed88c5d4d0 - Init START +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO comm 0xa2898e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x91ac8dce97e00f9a - Init START +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO comm 0x88d9810 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x91ac8dce97e00f9a - Init START +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO comm 0x89ce6e0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x91ac8dce97e00f9a - Init START +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO comm 0xa4a0280 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x91ac8dce97e00f9a - Init START +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO comm 0x9cfba00 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x91ac8dce97e00f9a - Init START +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO comm 0x8a5fb40 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x91ac8dce97e00f9a - Init START +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO comm 0x8f0c880 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x91ac8dce97e00f9a - Init START +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO comm 0x9e41a40 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x91ac8dce97e00f9a - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-142:114199:114527 [7] NCCL INFO Using network Libfabric +ip-26-0-172-142:114192:114526 [0] NCCL INFO Using network Libfabric +ip-26-0-172-142:114198:114529 [6] NCCL INFO Using network Libfabric +ip-26-0-172-142:114196:114530 [4] NCCL INFO Using network Libfabric +ip-26-0-172-142:114193:114531 [1] NCCL INFO Using network Libfabric +ip-26-0-172-142:114197:114528 [5] NCCL INFO Using network Libfabric +ip-26-0-172-142:114195:114532 [3] NCCL INFO Using network Libfabric +ip-26-0-172-142:114194:114533 [2] NCCL INFO Using network Libfabric +ip-26-0-172-142:114195:114532 [3] NCCL INFO comm 0x9ac31c0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc59569bdd2d5ffed - Init START +ip-26-0-172-142:114194:114533 [2] NCCL INFO comm 0x96b2190 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc59569bdd2d5ffed - Init START +ip-26-0-172-142:114196:114530 [4] NCCL INFO comm 0x8b9b650 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc59569bdd2d5ffed - Init START +ip-26-0-172-142:114197:114528 [5] NCCL INFO comm 0xa52a080 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc59569bdd2d5ffed - Init START +ip-26-0-172-142:114198:114529 [6] NCCL INFO comm 0xa0b4ab0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc59569bdd2d5ffed - Init START +ip-26-0-172-142:114199:114527 [7] NCCL INFO comm 0x9735ac0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc59569bdd2d5ffed - Init START +ip-26-0-172-142:114192:114526 [0] NCCL INFO comm 0x9b6d940 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc59569bdd2d5ffed - Init START +ip-26-0-172-142:114193:114531 [1] NCCL INFO comm 0x9bb1d50 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc59569bdd2d5ffed - Init START +ip-26-0-172-142:114198:114529 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114197:114528 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114196:114530 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114195:114532 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114193:114531 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114192:114526 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114199:114527 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114194:114533 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-142:114193:114531 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-142:114193:114531 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-142:114192:114526 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-142:114196:114530 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114196:114530 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-142:114199:114527 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114199:114527 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-142:114195:114532 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-142:114195:114532 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-142:114194:114533 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-142:114194:114533 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-142:114198:114529 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114198:114529 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-142:114197:114528 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114197:114528 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-142:114197:114528 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-172-142:114197:114528 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-172-142:114198:114529 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114199:114527 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-172-142:114196:114530 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114194:114533 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-172-142:114199:114527 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-142:114196:114530 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114193:114531 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114194:114533 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114193:114531 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-142:114195:114532 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114195:114532 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-172-142:114192:114526 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO NVLS multicast support is available on dev 6 +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO NVLS multicast support is available on dev 0 +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO NVLS multicast support is available on dev 1 +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO NVLS multicast support is available on dev 2 +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO NVLS multicast support is available on dev 5 +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO NVLS multicast support is available on dev 7 +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO NVLS multicast support is available on dev 4 +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO NVLS multicast support is available on dev 3 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 00/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 01/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 02/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 03/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 04/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Trees [0] -1/-1/-1->7->6 [1] -1/-1/-1->7->6 [2] -1/-1/-1->7->6 [3] -1/-1/-1->7->6 [4] -1/-1/-1->7->6 [5] -1/-1/-1->7->6 [6] -1/-1/-1->7->6 [7] -1/-1/-1->7->6 [8] -1/-1/-1->7->6 [9] -1/-1/-1->7->6 [10] -1/-1/-1->7->6 [11] -1/-1/-1->7->6 [12] -1/-1/-1->7->6 [13] -1/-1/-1->7->6 [14] -1/-1/-1->7->6 [15] -1/-1/-1->7->6 [16] -1/-1/-1->7->6 [17] -1/-1/-1->7->6 [18] -1/-1/-1->7->6 [19] -1/-1/-1->7->6 [20] -1/-1/-1->7->6 [21] -1/-1/-1->7->6 [22] -1/-1/-1->7->6 [23] -1/-1/-1->7->6 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 05/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1 [2] 3/-1/-1->2->1 [3] 3/-1/-1->2->1 [4] 3/-1/-1->2->1 [5] 3/-1/-1->2->1 [6] 3/-1/-1->2->1 [7] 3/-1/-1->2->1 [8] 3/-1/-1->2->1 [9] 3/-1/-1->2->1 [10] 3/-1/-1->2->1 [11] 3/-1/-1->2->1 [12] 3/-1/-1->2->1 [13] 3/-1/-1->2->1 [14] 3/-1/-1->2->1 [15] 3/-1/-1->2->1 [16] 3/-1/-1->2->1 [17] 3/-1/-1->2->1 [18] 3/-1/-1->2->1 [19] 3/-1/-1->2->1 [20] 3/-1/-1->2->1 [21] 3/-1/-1->2->1 [22] 3/-1/-1->2->1 [23] 3/-1/-1->2->1 +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0 [2] 2/-1/-1->1->0 [3] 2/-1/-1->1->0 [4] 2/-1/-1->1->0 [5] 2/-1/-1->1->0 [6] 2/-1/-1->1->0 [7] 2/-1/-1->1->0 [8] 2/-1/-1->1->0 [9] 2/-1/-1->1->0 [10] 2/-1/-1->1->0 [11] 2/-1/-1->1->0 [12] 2/-1/-1->1->0 [13] 2/-1/-1->1->0 [14] 2/-1/-1->1->0 [15] 2/-1/-1->1->0 [16] 2/-1/-1->1->0 [17] 2/-1/-1->1->0 [18] 2/-1/-1->1->0 [19] 2/-1/-1->1->0 [20] 2/-1/-1->1->0 [21] 2/-1/-1->1->0 [22] 2/-1/-1->1->0 [23] 2/-1/-1->1->0 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 06/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 07/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Trees [0] 4/-1/-1->3->2 [1] 4/-1/-1->3->2 [2] 4/-1/-1->3->2 [3] 4/-1/-1->3->2 [4] 4/-1/-1->3->2 [5] 4/-1/-1->3->2 [6] 4/-1/-1->3->2 [7] 4/-1/-1->3->2 [8] 4/-1/-1->3->2 [9] 4/-1/-1->3->2 [10] 4/-1/-1->3->2 [11] 4/-1/-1->3->2 [12] 4/-1/-1->3->2 [13] 4/-1/-1->3->2 [14] 4/-1/-1->3->2 [15] 4/-1/-1->3->2 [16] 4/-1/-1->3->2 [17] 4/-1/-1->3->2 [18] 4/-1/-1->3->2 [19] 4/-1/-1->3->2 [20] 4/-1/-1->3->2 [21] 4/-1/-1->3->2 [22] 4/-1/-1->3->2 [23] 4/-1/-1->3->2 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 08/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Trees [0] 5/-1/-1->4->3 [1] 5/-1/-1->4->3 [2] 5/-1/-1->4->3 [3] 5/-1/-1->4->3 [4] 5/-1/-1->4->3 [5] 5/-1/-1->4->3 [6] 5/-1/-1->4->3 [7] 5/-1/-1->4->3 [8] 5/-1/-1->4->3 [9] 5/-1/-1->4->3 [10] 5/-1/-1->4->3 [11] 5/-1/-1->4->3 [12] 5/-1/-1->4->3 [13] 5/-1/-1->4->3 [14] 5/-1/-1->4->3 [15] 5/-1/-1->4->3 [16] 5/-1/-1->4->3 [17] 5/-1/-1->4->3 [18] 5/-1/-1->4->3 [19] 5/-1/-1->4->3 [20] 5/-1/-1->4->3 [21] 5/-1/-1->4->3 [22] 5/-1/-1->4->3 [23] 5/-1/-1->4->3 +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 09/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Trees [0] 6/-1/-1->5->4 [1] 6/-1/-1->5->4 [2] 6/-1/-1->5->4 [3] 6/-1/-1->5->4 [4] 6/-1/-1->5->4 [5] 6/-1/-1->5->4 [6] 6/-1/-1->5->4 [7] 6/-1/-1->5->4 [8] 6/-1/-1->5->4 [9] 6/-1/-1->5->4 [10] 6/-1/-1->5->4 [11] 6/-1/-1->5->4 [12] 6/-1/-1->5->4 [13] 6/-1/-1->5->4 [14] 6/-1/-1->5->4 [15] 6/-1/-1->5->4 [16] 6/-1/-1->5->4 [17] 6/-1/-1->5->4 [18] 6/-1/-1->5->4 [19] 6/-1/-1->5->4 [20] 6/-1/-1->5->4 [21] 6/-1/-1->5->4 [22] 6/-1/-1->5->4 [23] 6/-1/-1->5->4 +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 10/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 11/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 12/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 13/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 14/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 15/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 16/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Trees [0] 7/-1/-1->6->5 [1] 7/-1/-1->6->5 [2] 7/-1/-1->6->5 [3] 7/-1/-1->6->5 [4] 7/-1/-1->6->5 [5] 7/-1/-1->6->5 [6] 7/-1/-1->6->5 [7] 7/-1/-1->6->5 [8] 7/-1/-1->6->5 [9] 7/-1/-1->6->5 [10] 7/-1/-1->6->5 [11] 7/-1/-1->6->5 [12] 7/-1/-1->6->5 [13] 7/-1/-1->6->5 [14] 7/-1/-1->6->5 [15] 7/-1/-1->6->5 [16] 7/-1/-1->6->5 [17] 7/-1/-1->6->5 [18] 7/-1/-1->6->5 [19] 7/-1/-1->6->5 [20] 7/-1/-1->6->5 [21] 7/-1/-1->6->5 [22] 7/-1/-1->6->5 [23] 7/-1/-1->6->5 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 17/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 18/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 19/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 20/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 21/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 22/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 23/24 : 0 1 2 3 4 5 6 7 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] 1/-1/-1->0->-1 [3] 1/-1/-1->0->-1 [4] 1/-1/-1->0->-1 [5] 1/-1/-1->0->-1 [6] 1/-1/-1->0->-1 [7] 1/-1/-1->0->-1 [8] 1/-1/-1->0->-1 [9] 1/-1/-1->0->-1 [10] 1/-1/-1->0->-1 [11] 1/-1/-1->0->-1 [12] 1/-1/-1->0->-1 [13] 1/-1/-1->0->-1 [14] 1/-1/-1->0->-1 [15] 1/-1/-1->0->-1 [16] 1/-1/-1->0->-1 [17] 1/-1/-1->0->-1 [18] 1/-1/-1->0->-1 [19] 1/-1/-1->0->-1 [20] 1/-1/-1->0->-1 [21] 1/-1/-1->0->-1 [22] 1/-1/-1->0->-1 [23] 1/-1/-1->0->-1 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO P2P Chunksize set to 524288 +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 00/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 01/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 00/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 00/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 02/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 01/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 01/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 03/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 02/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 02/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 04/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 03/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 04/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 03/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 05/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 04/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 00/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 04/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 05/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 04/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 06/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 05/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 01/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 06/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 07/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 05/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 06/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 05/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 02/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 07/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 03/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 08/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 08/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 04/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 09/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 10/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 07/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 06/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 06/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 11/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 08/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 07/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 07/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 00/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 12/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 09/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 08/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 08/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 01/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 13/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 10/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 09/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 09/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 02/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 09/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 14/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 11/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 05/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 10/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 10/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 03/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 10/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 15/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 12/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 06/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 11/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 11/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 04/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 11/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 16/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 13/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 07/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 12/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 12/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 05/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 12/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 17/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 14/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 08/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 13/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 13/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 06/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 13/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 18/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 15/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 09/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 14/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 14/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 07/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 14/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 19/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 16/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 10/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 15/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 15/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 08/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 15/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 20/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 17/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 11/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 04/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 16/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 09/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 16/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 16/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 21/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 18/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 12/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 05/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 17/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 10/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 17/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 19/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 22/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 17/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 13/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 06/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 18/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 11/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 18/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 23/0 : 1[1] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 18/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 20/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 14/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 19/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 07/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 12/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 19/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 19/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 15/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 21/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 20/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 08/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 13/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 20/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 20/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 16/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 22/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 09/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 21/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 14/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 21/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 21/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 17/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 23/0 : 2[2] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 22/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 10/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 15/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 22/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 22/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 18/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 23/0 : 6[6] -> 7[7] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 11/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 16/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 23/0 : 7[7] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 23/0 : 4[4] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 19/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 12/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 17/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 20/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 13/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 18/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 21/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 14/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 19/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 22/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 15/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 20/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 23/0 : 5[5] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 16/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 21/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 17/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 22/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 18/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 23/0 : 3[3] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 19/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 20/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 21/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 22/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Channel 23/0 : 0[0] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Connected all rings +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Connected all rings +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Connected all rings +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Connected all rings +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Connected all rings +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Connected all rings +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Connected all rings +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Connected all rings +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Connected all rings +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114192:114526 [0] NCCL INFO Connected all rings +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Connected all rings +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Connected all rings +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Connected all rings +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Connected all rings +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Connected all rings +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Connected all rings +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Connected all rings +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Connected all rings +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Connected all rings +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Connected all rings +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Connected all rings +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Connected all rings +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Connected all rings +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Connected all rings +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114198:114529 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Connected all rings +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114197:114528 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Connected all rings +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114195:114532 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Connected all rings +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Connected all rings +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114194:114533 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Connected all rings +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Connected all rings +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 00/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Connected all rings +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 01/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 02/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 03/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 04/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114193:114531 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-142:114196:114530 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 05/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 06/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-142:114199:114527 [7] NCCL INFO Connected all trees +ip-26-0-172-142:114199:114527 [7] NCCL INFO NVLS comm 0x9735ac0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 07/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 08/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 09/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 10/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 11/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 00/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 12/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 01/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 13/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 02/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 14/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 03/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 15/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 04/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 16/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 05/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 00/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 17/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 01/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 06/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 02/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 18/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 19/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 20/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 21/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 02/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 22/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 03/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 03/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Channel 23/0 : 7[7] -> 6[6] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 04/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 04/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 07/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 05/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 05/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 08/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 00/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 06/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 06/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 04/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Connected all trees +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 09/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO NVLS comm 0x9cfba00 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 01/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 07/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 07/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 05/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 10/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 02/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 08/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 08/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 06/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 11/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 04/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 03/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 09/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 09/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 07/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 10/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 08/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 12/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 04/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 11/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 09/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 13/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 05/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 12/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 10/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 14/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 06/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 13/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 11/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 15/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 07/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 14/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 12/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 16/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 08/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 15/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 13/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 17/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 09/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 14/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 16/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 18/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 10/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 15/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 17/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 19/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 11/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 05/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 16/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 18/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 10/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 12/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 20/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 17/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 19/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 13/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 21/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 18/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 20/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 14/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 22/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 19/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 21/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 06/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 15/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Channel 23/0 : 5[5] -> 4[4] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 20/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 22/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 07/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 16/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 21/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Channel 23/0 : 3[3] -> 2[2] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 08/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 17/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 11/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 22/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 09/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 18/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 12/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Channel 23/0 : 2[2] -> 1[1] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 10/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 19/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 13/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 11/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 20/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 14/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 12/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 21/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 15/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 13/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 16/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 22/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 14/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 17/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Channel 23/0 : 6[6] -> 5[5] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 15/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 18/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 16/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 19/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 17/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 20/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 18/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 21/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 19/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 22/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 20/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Connected all trees +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO NVLS comm 0xa2a71c0 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 21/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 22/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Channel 23/0 : 1[1] -> 0[0] via P2P/IPC +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Channel 23/0 : 4[4] -> 3[3] via P2P/IPC +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Connected all trees +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO NVLS comm 0x976c800 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114198:114529 [6] NCCL INFO Connected all trees +ip-26-0-172-142:114198:114529 [6] NCCL INFO NVLS comm 0xa0b4ab0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Connected all trees +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO NVLS comm 0x9541bd0 headRank 7 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Connected all trees +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO NVLS comm 0xa4a0280 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Connected all trees +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO NVLS comm 0xa594020 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114192:114526 [0] NCCL INFO Connected all trees +ip-26-0-172-142:114192:114526 [0] NCCL INFO NVLS comm 0x9b6d940 headRank 0 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Connected all trees +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO NVLS comm 0xa59a7e0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Connected all trees +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO NVLS comm 0x9b66f50 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Connected all trees +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO NVLS comm 0x88d9810 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Connected all trees +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO NVLS comm 0x9e41a40 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Connected all trees +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO NVLS comm 0x8f0c880 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Connected all trees +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Connected all trees +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO NVLS comm 0x963d1d0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO NVLS comm 0x8a5fb40 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114194:114533 [2] NCCL INFO Connected all trees +ip-26-0-172-142:114194:114533 [2] NCCL INFO NVLS comm 0x96b2190 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Connected all trees +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO NVLS comm 0x89ce6e0 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114193:114531 [1] NCCL INFO Connected all trees +ip-26-0-172-142:114193:114531 [1] NCCL INFO NVLS comm 0x9bb1d50 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114197:114528 [5] NCCL INFO Connected all trees +ip-26-0-172-142:114197:114528 [5] NCCL INFO NVLS comm 0xa52a080 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Connected all trees +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO NVLS comm 0xa2898e0 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Connected all trees +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO NVLS comm 0x9df3bc0 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114196:114530 [4] NCCL INFO Connected all trees +ip-26-0-172-142:114196:114530 [4] NCCL INFO NVLS comm 0x8b9b650 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114195:114532 [3] NCCL INFO Connected all trees +ip-26-0-172-142:114195:114532 [3] NCCL INFO NVLS comm 0x9ac31c0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Connected all trees +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO NVLS comm 0x95e7530 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Connected all trees +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO NVLS comm 0x8895220 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Connected all trees +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO NVLS comm 0xa5c5ab0 headRank 6 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Connected all trees +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO NVLS comm 0x93e3960 headRank 1 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114193:114531 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114193:114531 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114193:114531 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-142:114193:114531 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114199:114527 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114199:114527 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114199:114527 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-142:114199:114527 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114195:114532 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114195:114532 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114195:114532 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-142:114195:114532 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114192:114526 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114192:114526 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114198:114529 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114198:114529 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114198:114529 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-142:114198:114529 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114192:114526 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-142:114192:114526 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114196:114530 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-142:114196:114530 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-142:114196:114530 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-142:114196:114530 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114194:114533 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114194:114533 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-142:114194:114533 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-142:114194:114533 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114197:114528 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114197:114528 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114197:114528 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-142:114197:114528 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Connected all trees +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO NVLS comm 0x94bea50 headRank 2 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Connected all trees +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO NVLS comm 0xa0628c0 headRank 4 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-142:114196:114530 [4] NCCL INFO comm 0x8b9b650 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc59569bdd2d5ffed - Init COMPLETE +ip-26-0-172-142:114198:114529 [6] NCCL INFO comm 0xa0b4ab0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xc59569bdd2d5ffed - Init COMPLETE +ip-26-0-172-142:114192:114526 [0] NCCL INFO comm 0x9b6d940 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xc59569bdd2d5ffed - Init COMPLETE +ip-26-0-172-142:114194:114533 [2] NCCL INFO comm 0x96b2190 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xc59569bdd2d5ffed - Init COMPLETE +ip-26-0-172-142:114195:114532 [3] NCCL INFO comm 0x9ac31c0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xc59569bdd2d5ffed - Init COMPLETE +ip-26-0-172-142:114197:114528 [5] NCCL INFO comm 0xa52a080 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xc59569bdd2d5ffed - Init COMPLETE +ip-26-0-172-142:114199:114527 [7] NCCL INFO comm 0x9735ac0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xc59569bdd2d5ffed - Init COMPLETE +ip-26-0-172-142:114193:114531 [1] NCCL INFO comm 0x9bb1d50 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc59569bdd2d5ffed - Init COMPLETE +ip-26-0-172-73:2574289:2574617 [7] NCCL INFO comm 0x9cfba00 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x91ac8dce97e00f9a - Init COMPLETE +ip-26-0-172-73:2574287:2574620 [5] NCCL INFO comm 0x9e41a40 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x91ac8dce97e00f9a - Init COMPLETE +ip-26-0-172-73:2574283:2574622 [1] NCCL INFO comm 0x89ce6e0 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x91ac8dce97e00f9a - Init COMPLETE +ip-26-0-172-73:2574288:2574619 [6] NCCL INFO comm 0x8a5fb40 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x91ac8dce97e00f9a - Init COMPLETE +ip-26-0-172-73:2574285:2574618 [3] NCCL INFO comm 0x88d9810 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x91ac8dce97e00f9a - Init COMPLETE +ip-26-0-172-73:2574282:2574616 [0] NCCL INFO comm 0xa4a0280 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x91ac8dce97e00f9a - Init COMPLETE +ip-26-0-172-73:2574286:2574621 [4] NCCL INFO comm 0x8f0c880 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x91ac8dce97e00f9a - Init COMPLETE +ip-26-0-172-73:2574284:2574623 [2] NCCL INFO comm 0xa2898e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x91ac8dce97e00f9a - Init COMPLETE +ip-26-0-172-147:2713674:2714009 [1] NCCL INFO comm 0x9b66f50 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0x7c40d7ed88c5d4d0 - Init COMPLETE +ip-26-0-172-147:2713678:2714014 [5] NCCL INFO comm 0x9df3bc0 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7c40d7ed88c5d4d0 - Init COMPLETE +ip-26-0-172-147:2713680:2714013 [7] NCCL INFO comm 0x976c800 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0x7c40d7ed88c5d4d0 - Init COMPLETE +ip-26-0-172-147:2713679:2714010 [6] NCCL INFO comm 0x963d1d0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7c40d7ed88c5d4d0 - Init COMPLETE +ip-26-0-172-147:2713676:2714011 [3] NCCL INFO comm 0x8895220 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7c40d7ed88c5d4d0 - Init COMPLETE +ip-26-0-172-147:2713675:2714012 [2] NCCL INFO comm 0xa59a7e0 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0x7c40d7ed88c5d4d0 - Init COMPLETE +ip-26-0-172-147:2713677:2714008 [4] NCCL INFO comm 0x95e7530 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0x7c40d7ed88c5d4d0 - Init COMPLETE +ip-26-0-172-147:2713673:2714007 [0] NCCL INFO comm 0xa2a71c0 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0x7c40d7ed88c5d4d0 - Init COMPLETE +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Connected all trees +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO NVLS comm 0x9a056a0 headRank 3 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Connected all trees +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO NVLS comm 0x99c3660 headRank 5 nHeads 8 buffSize 4194304 memSize 2097152 nvlsPerRankSize 201326592 nvlsTotalSize 1610612736 +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO Connected NVLS tree +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO threadThresholds 8/8/64 | 64/8/64 | 512 | 512 +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO 24 coll channels, 16 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888252:3888593 [4] NCCL INFO comm 0xa0628c0 rank 4 nranks 8 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb335559d1e773894 - Init COMPLETE +ip-26-0-172-116:3888250:3888588 [2] NCCL INFO comm 0x94bea50 rank 2 nranks 8 cudaDev 2 nvmlDev 2 busId 75000 commId 0xb335559d1e773894 - Init COMPLETE +ip-26-0-172-116:3888248:3888586 [0] NCCL INFO comm 0xa594020 rank 0 nranks 8 cudaDev 0 nvmlDev 0 busId 53000 commId 0xb335559d1e773894 - Init COMPLETE +ip-26-0-172-116:3888254:3888592 [6] NCCL INFO comm 0xa5c5ab0 rank 6 nranks 8 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb335559d1e773894 - Init COMPLETE +ip-26-0-172-116:3888255:3888587 [7] NCCL INFO comm 0x9541bd0 rank 7 nranks 8 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb335559d1e773894 - Init COMPLETE +ip-26-0-172-116:3888253:3888589 [5] NCCL INFO comm 0x99c3660 rank 5 nranks 8 cudaDev 5 nvmlDev 5 busId a8000 commId 0xb335559d1e773894 - Init COMPLETE +ip-26-0-172-116:3888249:3888591 [1] NCCL INFO comm 0x93e3960 rank 1 nranks 8 cudaDev 1 nvmlDev 1 busId 64000 commId 0xb335559d1e773894 - Init COMPLETE +ip-26-0-172-116:3888251:3888590 [3] NCCL INFO comm 0x9a056a0 rank 3 nranks 8 cudaDev 3 nvmlDev 3 busId 86000 commId 0xb335559d1e773894 - Init COMPLETE +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Config: +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Config(general=GeneralArgs(project='debug', +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: run='8.86G_dp8_tp1_pp4_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k', +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: seed=42, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: step=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: consumed_train_samples=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: ignore_sanity_checks=True), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: parallelism=ParallelismArgs(dp=8, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: pp=4, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: tp=1, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: pp_engine=, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: tp_mode=, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: tp_linear_async_communication=True, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: recompute_layer=False, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: tp_recompute_allgather=True, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: expert_parallel_size=1), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: eos_token_id=0, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: hidden_act='silu', +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: hidden_size=4096, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: initializer_range=0.02, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: intermediate_size=14336, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: is_llama_config=True, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: max_position_embeddings=4096, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: num_attention_heads=32, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: num_hidden_layers=32, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: num_key_value_heads=32, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: pad_token_id=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: pretraining_tp=1, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: rms_norm_eps=1e-05, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: rope_scaling=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: rope_theta=10000.0, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: rope_interleaved=False, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: tie_word_embeddings=False, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: use_cache=True, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: vocab_size=131072), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: init_method=RandomInit(std=0.02), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: dtype=torch.bfloat16, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: make_vocab_size_divisible_by=1, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: ddp_bucket_cap_mb=25), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: tokenizer_revision=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: tokenizer_max_length=None), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: checkpoint_interval=10000, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: save_initial_state=False, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: save_final_state=False, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: resume_checkpoint_path=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: checkpoints_path_is_shared_file_system=False), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: logging=LoggingArgs(log_level='info', +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: log_level_replica='info', +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: iteration_step_info_interval=1), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: tokens=TokensArgs(sequence_length=4096, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: train_steps=100, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: micro_batch_size=1, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: batch_accumulation_per_replica=32, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: val_check_interval=100, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: limit_val_batches=0, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: limit_test_batches=0), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: adam_beta1=0.9, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: adam_beta2=0.95, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: torch_adam_is_fused=True, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: name='adamW'), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: zero_stage=1, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: weight_decay=0.01, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: clip_grad=1.0, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: accumulate_grad_in_fp32=True, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: lr_warmup_steps=2, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: lr_warmup_style='linear', +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: lr_decay_style='cosine', +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: lr_decay_steps=13, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: lr_decay_starting_step=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: min_decay_lr=1e-05)), +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: start_training_step=1, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: data=DataArgs(dataset=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: seed=42, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: num_loading_workers=1))], +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: profiler=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: lighteval=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: s3_upload=None) +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Model Config: +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: LlamaConfig(bos_token_id=0, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: eos_token_id=0, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: hidden_act='silu', +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: hidden_size=4096, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: initializer_range=0.02, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: intermediate_size=14336, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: is_llama_config=True, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: max_position_embeddings=4096, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: num_attention_heads=32, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: num_hidden_layers=32, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: num_key_value_heads=32, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: pad_token_id=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: pretraining_tp=1, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: rms_norm_eps=1e-05, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: rope_scaling=None, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: rope_theta=10000.0, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: rope_interleaved=False, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: tie_word_embeddings=False, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: use_cache=True, +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: vocab_size=131072) +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Building model.. +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Initialize RoPE Theta = 10000.0 +12/28/2024 02:08:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Using network Libfabric +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO comm 0x9bad6b0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x10ec4f5d7c5178c6 - Init START +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO comm 0xb113310 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7ecea155d2a74d67 - Init START +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO comm 0xb0cbb10 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa15433dbf4a705c1 - Init START +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO comm 0x9d2f7b0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1a30be8e5ae5d5a1 - Init START +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO comm 0x9ca1ba0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeb451c5f94c7086c - Init START +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO comm 0xb775440 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4a3244410bdf9af4 - Init START +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO comm 0xafcd1c0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb0ff441d4eadfdc7 - Init START +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO comm 0xb558db0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3e9a0b1b4f9f88eb - Init START +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO comm 0xaa44b50 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xca39b8b72fcf2b03 - Init START +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +NCCL version 2.18.5+cuda12.2 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO comm 0xa1dbce0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x74d7505d5e3708b8 - Init START +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO comm 0xba2d600 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4ff999083ee29086 - Init START +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO comm 0xae40530 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf9cd9325b8b373f4 - Init START +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO comm 0xa8bde30 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x3a21f894d932dede - Init START +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO comm 0xa913470 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x42856517f918fe40 - Init START +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO comm 0x9b6d060 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x404023b69a21d2d6 - Init START +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Using network Libfabric +ip-26-0-172-142:114196:114568 [4] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Using network Libfabric +ip-26-0-172-142:114198:114564 [6] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Using network Libfabric +ip-26-0-172-142:114193:114571 [1] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Using network Libfabric +ip-26-0-172-142:114199:114566 [7] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO comm 0xa8805c0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc355da503c9ccffd - Init START +ip-26-0-172-142:114195:114575 [3] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Using network Libfabric +ip-26-0-172-142:114196:114568 [4] NCCL INFO comm 0x9e72700 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc37ecc58a37a30e8 - Init START +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO comm 0xa9595c0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xabeda441492e49de - Init START +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114198:114564 [6] NCCL INFO comm 0xb38cfd0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb0e277f560703498 - Init START +ip-26-0-172-142:114192:114573 [0] NCCL INFO Using network Libfabric +ip-26-0-172-142:114194:114576 [2] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO comm 0xb581140 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf291f9b01d21d559 - Init START +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO comm 0xb4f9ea0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2059e7085330dcb1 - Init START +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Using network Libfabric +ip-26-0-172-142:114197:114574 [5] NCCL INFO Using network Libfabric +ip-26-0-172-142:114196:114568 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114193:114571 [1] NCCL INFO comm 0xae8a4c0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x99233f0534ff668d - Init START +ip-26-0-172-142:114198:114564 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO comm 0xae5ac50 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe2745285a347fe7d - Init START +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114199:114566 [7] NCCL INFO comm 0xaa0ede0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcb6118fa37764a95 - Init START +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO comm 0xba5e040 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x60fb6349f734c6c9 - Init START +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO comm 0xb870f30 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5a0d96e0370f8a70 - Init START +ip-26-0-172-142:114193:114571 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114195:114575 [3] NCCL INFO comm 0xb7e0f10 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa39cab674b29038 - Init START +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO comm 0xa9d9430 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xeca4621286544925 - Init START +ip-26-0-172-142:114199:114566 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114194:114576 [2] NCCL INFO comm 0xa98a260 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8f42d18d8b64b622 - Init START +ip-26-0-172-142:114192:114573 [0] NCCL INFO comm 0xb88ca90 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8b70cd65afb9b1b8 - Init START +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114195:114575 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114197:114574 [5] NCCL INFO comm 0xb8022f0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe38295340a16dbab - Init START +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO comm 0xae9f7a0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7d751067f36a2b9d - Init START +ip-26-0-172-142:114194:114576 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114192:114573 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114197:114574 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Connected all rings +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO Connected all trees +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Connected all rings +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO Connected all trees +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Connected all rings +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO Connected all trees +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713678:2714045 [5] NCCL INFO comm 0xb0cbb10 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa15433dbf4a705c1 - Init COMPLETE +ip-26-0-172-142:114198:114564 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-142:114198:114564 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114198:114564 [6] NCCL INFO Connected all rings +ip-26-0-172-142:114198:114564 [6] NCCL INFO Connected all trees +ip-26-0-172-142:114198:114564 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Connected all rings +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO Connected all trees +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Connected all rings +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO Connected all trees +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Connected all rings +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO Connected all trees +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Connected all rings +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO Connected all trees +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Connected all rings +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO Connected all trees +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114195:114575 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-142:114195:114575 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114195:114575 [3] NCCL INFO Connected all rings +ip-26-0-172-142:114195:114575 [3] NCCL INFO Connected all trees +ip-26-0-172-142:114195:114575 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574288:2574661 [6] NCCL INFO comm 0x9d2f7b0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x1a30be8e5ae5d5a1 - Init COMPLETE +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO Connected all trees +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713676:2714054 [3] NCCL INFO comm 0x9b6d060 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x404023b69a21d2d6 - Init COMPLETE +ip-26-0-172-147:2713679:2714049 [6] NCCL INFO comm 0xa913470 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x42856517f918fe40 - Init COMPLETE +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-73:2574287:2574664 [5] NCCL INFO comm 0xb113310 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0x7ecea155d2a74d67 - Init COMPLETE +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Connected all rings +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO Connected all trees +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Connected all rings +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO Connected all trees +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888248:3888620 [0] NCCL INFO comm 0xba2d600 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4ff999083ee29086 - Init COMPLETE +ip-26-0-172-142:114198:114564 [6] NCCL INFO comm 0xb38cfd0 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0xb0e277f560703498 - Init COMPLETE +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Connected all rings +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO Connected all trees +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888254:3888631 [6] NCCL INFO comm 0xba5e040 rank 0 nranks 1 cudaDev 6 nvmlDev 6 busId b9000 commId 0x60fb6349f734c6c9 - Init COMPLETE +ip-26-0-172-116:3888255:3888633 [7] NCCL INFO comm 0xa9d9430 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xeca4621286544925 - Init COMPLETE +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Connected all rings +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO Connected all trees +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574282:2574662 [0] NCCL INFO comm 0xb775440 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x4a3244410bdf9af4 - Init COMPLETE +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Using network Libfabric +ip-26-0-172-142:114198:114582 [6] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-142:114195:114575 [3] NCCL INFO comm 0xb7e0f10 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0xfa39cab674b29038 - Init COMPLETE +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO comm 0xa7745e0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2d469fc5421fa358 - Init START +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO comm 0xc4a2c10 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2d469fc5421fa358 - Init START +ip-26-0-172-142:114198:114582 [6] NCCL INFO comm 0xbdd1ec0 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2d469fc5421fa358 - Init START +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO comm 0xb3584a0 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2d469fc5421fa358 - Init START +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114198:114582 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Connected all rings +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO Connected all trees +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Connected all rings +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO Connected all trees +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Connected all rings +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO Connected all trees +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Connected all rings +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO Connected all trees +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Connected all rings +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO Connected all trees +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574289:2574663 [7] NCCL INFO comm 0xafcd1c0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xb0ff441d4eadfdc7 - Init COMPLETE +ip-26-0-172-73:2574285:2574659 [3] NCCL INFO comm 0x9bad6b0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x10ec4f5d7c5178c6 - Init COMPLETE +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Connected all rings +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO Connected all trees +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Connected all rings +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO Connected all trees +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Connected all rings +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO Connected all trees +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Connected all rings +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO Connected all trees +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-147:2713675:2714055 [2] NCCL INFO comm 0xb870f30 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x5a0d96e0370f8a70 - Init COMPLETE +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Connected all rings +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO Connected all trees +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-73:2574283:2574658 [1] NCCL INFO comm 0x9ca1ba0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xeb451c5f94c7086c - Init COMPLETE +ip-26-0-172-73:2574286:2574665 [4] NCCL INFO comm 0xa1dbce0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x74d7505d5e3708b8 - Init COMPLETE +ip-26-0-172-116:3888251:3888634 [3] NCCL INFO comm 0xae9f7a0 rank 0 nranks 1 cudaDev 3 nvmlDev 3 busId 86000 commId 0x7d751067f36a2b9d - Init COMPLETE +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Using network Libfabric +ip-26-0-172-142:114195:114583 [3] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Connected all rings +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO Connected all trees +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-147:2713680:2714047 [7] NCCL INFO comm 0xaa44b50 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xca39b8b72fcf2b03 - Init COMPLETE +ip-26-0-172-147:2713674:2714050 [1] NCCL INFO comm 0xae40530 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf9cd9325b8b373f4 - Init COMPLETE +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO comm 0xa5f2710 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2cba1ebe76552e9e - Init START +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO comm 0xb8e3fa0 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2cba1ebe76552e9e - Init START +ip-26-0-172-142:114195:114583 [3] NCCL INFO comm 0xb7ebf10 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2cba1ebe76552e9e - Init START +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO comm 0xa5b21f0 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2cba1ebe76552e9e - Init START +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114195:114583 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574284:2574666 [2] NCCL INFO comm 0xb558db0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3e9a0b1b4f9f88eb - Init COMPLETE +ip-26-0-172-147:2713677:2714051 [4] NCCL INFO comm 0xa8bde30 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x3a21f894d932dede - Init COMPLETE +ip-26-0-172-116:3888252:3888628 [4] NCCL INFO comm 0xb4f9ea0 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2059e7085330dcb1 - Init COMPLETE +ip-26-0-172-116:3888250:3888629 [2] NCCL INFO comm 0xa9595c0 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0xabeda441492e49de - Init COMPLETE +ip-26-0-172-142:114192:114573 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-142:114192:114573 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114192:114573 [0] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114573 [0] NCCL INFO Connected all trees +ip-26-0-172-142:114192:114573 [0] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888249:3888627 [1] NCCL INFO comm 0xa8805c0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0xc355da503c9ccffd - Init COMPLETE +ip-26-0-172-147:2713673:2714057 [0] NCCL INFO comm 0xb581140 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0xf291f9b01d21d559 - Init COMPLETE +ip-26-0-172-142:114199:114566 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-142:114199:114566 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114199:114566 [7] NCCL INFO Connected all rings +ip-26-0-172-142:114199:114566 [7] NCCL INFO Connected all trees +ip-26-0-172-142:114199:114566 [7] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114193:114571 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-142:114193:114571 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-142:114193:114571 [1] NCCL INFO Connected all rings +ip-26-0-172-142:114193:114571 [1] NCCL INFO Connected all trees +ip-26-0-172-142:114193:114571 [1] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114197:114574 [5] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-142:114197:114574 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114197:114574 [5] NCCL INFO Connected all rings +ip-26-0-172-142:114197:114574 [5] NCCL INFO Connected all trees +ip-26-0-172-142:114197:114574 [5] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-116:3888253:3888630 [5] NCCL INFO comm 0xae5ac50 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe2745285a347fe7d - Init COMPLETE +ip-26-0-172-142:114194:114576 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-142:114194:114576 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 00/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 01/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 02/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 03/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 04/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 05/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 06/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 07/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 08/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 09/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 10/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 11/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 12/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 13/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 14/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 15/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 16/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 17/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 18/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 19/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 20/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 21/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 22/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 23/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 24/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 25/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 26/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 27/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 28/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 29/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 30/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Channel 31/32 : 0 +ip-26-0-172-142:114196:114568 [4] NCCL INFO Trees [0] -1/-1/-1->0->-1 [1] -1/-1/-1->0->-1 [2] -1/-1/-1->0->-1 [3] -1/-1/-1->0->-1 [4] -1/-1/-1->0->-1 [5] -1/-1/-1->0->-1 [6] -1/-1/-1->0->-1 [7] -1/-1/-1->0->-1 [8] -1/-1/-1->0->-1 [9] -1/-1/-1->0->-1 [10] -1/-1/-1->0->-1 [11] -1/-1/-1->0->-1 [12] -1/-1/-1->0->-1 [13] -1/-1/-1->0->-1 [14] -1/-1/-1->0->-1 [15] -1/-1/-1->0->-1 [16] -1/-1/-1->0->-1 [17] -1/-1/-1->0->-1 [18] -1/-1/-1->0->-1 [19] -1/-1/-1->0->-1 [20] -1/-1/-1->0->-1 [21] -1/-1/-1->0->-1 [22] -1/-1/-1->0->-1 [23] -1/-1/-1->0->-1 [24] -1/-1/-1->0->-1 [25] -1/-1/-1->0->-1 [26] -1/-1/-1->0->-1 [27] -1/-1/-1->0->-1 [28] -1/-1/-1->0->-1 [29] -1/-1/-1->0->-1 [30] -1/-1/-1->0->-1 [31] -1/-1/-1->0->-1 +ip-26-0-172-142:114196:114568 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114194:114576 [2] NCCL INFO Connected all rings +ip-26-0-172-142:114194:114576 [2] NCCL INFO Connected all trees +ip-26-0-172-142:114194:114576 [2] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114196:114568 [4] NCCL INFO Connected all rings +ip-26-0-172-142:114196:114568 [4] NCCL INFO Connected all trees +ip-26-0-172-142:114196:114568 [4] NCCL INFO 32 coll channels, 0 nvls channels, 32 p2p channels, 32 p2p channels per peer +ip-26-0-172-142:114192:114573 [0] NCCL INFO comm 0xb88ca90 rank 0 nranks 1 cudaDev 0 nvmlDev 0 busId 53000 commId 0x8b70cd65afb9b1b8 - Init COMPLETE +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Using network Libfabric +ip-26-0-172-142:114192:114596 [0] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO comm 0xc1ba6c0 rank 3 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x95b80b134b9b29b0 - Init START +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO comm 0xc471eb0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x95b80b134b9b29b0 - Init START +ip-26-0-172-142:114192:114596 [0] NCCL INFO comm 0xb897c90 rank 1 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x95b80b134b9b29b0 - Init START +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO comm 0xbfc6260 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x95b80b134b9b29b0 - Init START +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-172-142:114192:114596 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114199:114566 [7] NCCL INFO comm 0xaa0ede0 rank 0 nranks 1 cudaDev 7 nvmlDev 7 busId ca000 commId 0xcb6118fa37764a95 - Init COMPLETE +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Using network Libfabric +ip-26-0-172-142:114199:114597 [7] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO comm 0xba11fa0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x31dbc05cf868d21c - Init START +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO comm 0xb41e000 rank 0 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x31dbc05cf868d21c - Init START +ip-26-0-172-142:114199:114597 [7] NCCL INFO comm 0xb4538c0 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x31dbc05cf868d21c - Init START +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO comm 0xb489be0 rank 2 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x31dbc05cf868d21c - Init START +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114199:114597 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114193:114571 [1] NCCL INFO comm 0xae8a4c0 rank 0 nranks 1 cudaDev 1 nvmlDev 1 busId 64000 commId 0x99233f0534ff668d - Init COMPLETE +ip-26-0-172-142:114197:114574 [5] NCCL INFO comm 0xb8022f0 rank 0 nranks 1 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe38295340a16dbab - Init COMPLETE +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Using network Libfabric +ip-26-0-172-142:114193:114598 [1] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO comm 0xa6e6a00 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e7aade2168ed94a - Init START +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO comm 0xb2c5230 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e7aade2168ed94a - Init START +ip-26-0-172-142:114193:114598 [1] NCCL INFO comm 0xb8cf2a0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e7aade2168ed94a - Init START +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO comm 0xb885500 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e7aade2168ed94a - Init START +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114193:114598 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Using network Libfabric +ip-26-0-172-142:114197:114599 [5] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Using network Libfabric +ip-26-0-172-142:114194:114576 [2] NCCL INFO comm 0xa98a260 rank 0 nranks 1 cudaDev 2 nvmlDev 2 busId 75000 commId 0x8f42d18d8b64b622 - Init COMPLETE +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO comm 0xbb58200 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x32c4b1f092332bba - Init START +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO comm 0xb89f850 rank 0 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x32c4b1f092332bba - Init START +ip-26-0-172-142:114197:114599 [5] NCCL INFO comm 0xc247290 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x32c4b1f092332bba - Init START +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO comm 0xbb10150 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x32c4b1f092332bba - Init START +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114197:114599 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Using network Libfabric +ip-26-0-172-142:114194:114600 [2] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO comm 0xbf9de60 rank 3 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3ca14ff2825f2b5f - Init START +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO comm 0xb39e3f0 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3ca14ff2825f2b5f - Init START +ip-26-0-172-142:114194:114600 [2] NCCL INFO comm 0xb3cf0a0 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3ca14ff2825f2b5f - Init START +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO comm 0xc2b5df0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3ca14ff2825f2b5f - Init START +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114194:114600 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114196:114568 [4] NCCL INFO comm 0x9e72700 rank 0 nranks 1 cudaDev 4 nvmlDev 4 busId 97000 commId 0xc37ecc58a37a30e8 - Init COMPLETE +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Using network Libfabric +ip-26-0-172-142:114196:114601 [4] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO comm 0xac20ea0 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x320345d6074aeb37 - Init START +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO comm 0xbf3e4d0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x320345d6074aeb37 - Init START +ip-26-0-172-142:114196:114601 [4] NCCL INFO comm 0xa8b77a0 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x320345d6074aeb37 - Init START +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO comm 0xb302cb0 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x320345d6074aeb37 - Init START +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114196:114601 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-142:114195:114583 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-142:114195:114583 [3] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-172-142:114195:114583 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114194:114600 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-142:114198:114582 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114198:114582 [6] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-172-142:114198:114582 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114193:114598 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-142:114193:114598 [1] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-172-142:114193:114598 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114194:114600 [2] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-172-142:114194:114600 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-172-142:114196:114601 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 00/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 01/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 02/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 03/0 : 1[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 00/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 01/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 02/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 03/0 : 2[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 00/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 01/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 02/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 03/0 : 3[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-142:114192:114596 [0] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-172-142:114192:114596 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114197:114599 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114199:114597 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114197:114599 [5] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-172-142:114197:114599 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114199:114597 [7] NCCL INFO Trees [0] -1/-1/-1->1->2 [1] -1/-1/-1->1->2 [2] 2/0/-1->1->3 [3] 2/0/-1->1->3 +ip-26-0-172-142:114199:114597 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Trees [0] 1/3/-1->2->0 [1] 1/3/-1->2->0 [2] -1/-1/-1->2->1 [3] -1/-1/-1->2->1 +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2 [2] 1/-1/-1->3->-1 [3] 1/-1/-1->3->-1 +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 00/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 01/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 02/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 03/0 : 1[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 00/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 01/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 02/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 03/0 : 2[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 00/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 01/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 02/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 00/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 03/0 : 3[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 01/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 02/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 03/0 : 1[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 00/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 01/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 02/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 03/0 : 2[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 00/04 : 0 1 2 3 +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 01/04 : 0 1 2 3 +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 02/04 : 0 1 2 3 +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 03/04 : 0 1 2 3 +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Trees [0] 2/-1/-1->0->-1 [1] 2/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 00/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 01/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 02/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 03/0 : 3[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 00/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 01/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 02/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 03/0 : 3[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 00/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 01/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 02/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 03/0 : 3[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 00/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 01/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 02/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 03/0 : 1[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 00/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 01/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 02/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 03/0 : 2[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 00/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 01/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 02/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 03/0 : 3[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 00/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 01/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 02/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 03/0 : 1[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 00/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 01/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 02/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 03/0 : 2[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 00/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 01/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 02/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 03/0 : 1[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 00/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 01/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 02/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 03/0 : 3[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 00/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 01/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 02/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 03/0 : 3[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 00/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 01/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 02/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 03/0 : 3[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 00/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 01/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 02/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 03/0 : 3[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 00/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 00/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 01/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 01/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 02/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 03/0 : 2[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 02/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 03/0 : 1[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 00/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 01/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 02/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 03/0 : 3[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 00/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 00/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 01/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 02/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 01/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 03/0 : 2[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 02/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 00/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 03/0 : 1[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 01/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 00/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 02/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 01/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 03/0 : 3[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 02/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 03/0 : 2[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 00/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 01/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 02/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 03/0 : 1[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 00/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 01/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 02/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 03/0 : 2[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Connected all rings +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Connected all rings +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Connected all rings +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 02/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 03/0 : 3[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 02/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 03/0 : 1[3] -> 3[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114583 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Connected all rings +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 00/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 01/0 : 0[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 00/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 01/0 : 2[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 00/0 : 3[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 01/0 : 3[3] -> 2[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 00/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 01/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 02/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Channel 03/0 : 2[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Connected all rings +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Connected all rings +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Connected all rings +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Connected all rings +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Connected all rings +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 00/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 01/0 : 0[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 00/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 01/0 : 2[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 00/0 : 3[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 01/0 : 3[1] -> 2[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Connected all rings +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Connected all rings +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 00/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 01/0 : 0[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 00/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 01/0 : 2[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Connected all rings +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 02/0 : 3[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 03/0 : 3[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Connected all rings +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Connected all rings +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 02/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 03/0 : 3[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 02/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 03/0 : 1[1] -> 3[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 00/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 00/0 : 3[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 01/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Channel 01/0 : 3[1] -> 2[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 02/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 03/0 : 2[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114598 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Connected all rings +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Connected all rings +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 02/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 03/0 : 3[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 02/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 03/0 : 1[6] -> 3[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 00/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 00/0 : 3[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 01/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Channel 01/0 : 3[6] -> 2[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 02/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 03/0 : 2[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114582 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Connected all rings +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 00/0 : 2[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 01/0 : 2[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 00/0 : 0[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 01/0 : 0[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Connected all rings +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 00/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 01/0 : 0[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 00/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 01/0 : 2[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 00/0 : 3[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 01/0 : 3[2] -> 2[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Connected all rings +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 02/0 : 3[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 03/0 : 3[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 02/0 : 1[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 03/0 : 1[2] -> 3[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 00/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 00/0 : 3[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 01/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Channel 01/0 : 3[2] -> 2[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 02/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 03/0 : 2[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114600 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 02/0 : 1[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 03/0 : 1[0] -> 3[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 02/0 : 3[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 03/0 : 3[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Connected all rings +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Connected all rings +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Connected all rings +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Connected all rings +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 02/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 03/0 : 3[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 02/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 03/0 : 1[4] -> 3[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114601 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Connected all rings +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 00/0 : 2[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 01/0 : 2[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 00/0 : 0[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 01/0 : 0[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Connected all rings +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 00/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 01/0 : 0[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 00/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 01/0 : 2[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 00/0 : 3[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 01/0 : 3[4] -> 2[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 00/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 01/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 02/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Channel 03/0 : 2[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO Connected all trees +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Connected all rings +ip-26-0-172-142:114197:114599 [5] NCCL INFO Connected all rings +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 00/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 02/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 01/0 : 0[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 03/0 : 3[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 02/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 00/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 03/0 : 1[5] -> 3[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 01/0 : 2[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Connected all rings +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 00/0 : 3[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 01/0 : 3[0] -> 2[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 00/0 : 0[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 01/0 : 0[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114599 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 00/0 : 2[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 01/0 : 2[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Connected all rings +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 00/0 : 0[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 01/0 : 0[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 00/0 : 2[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 00/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 00/0 : 3[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 01/0 : 2[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 01/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Channel 01/0 : 3[5] -> 2[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 02/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Channel 03/0 : 2[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 00/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 01/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 02/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Channel 03/0 : 2[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574285:2574684 [3] NCCL INFO comm 0xa5f2710 rank 3 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2cba1ebe76552e9e - Init COMPLETE +ip-26-0-172-142:114199:114597 [7] NCCL INFO Connected all rings +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 02/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 03/0 : 3[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 02/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 03/0 : 1[7] -> 3[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 00/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 00/0 : 3[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 01/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Channel 01/0 : 3[7] -> 2[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 02/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 03/0 : 2[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114597 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 02/0 : 3[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 03/0 : 3[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 02/0 : 1[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 03/0 : 1[0] -> 3[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 00/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 01/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 02/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 03/0 : 2[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114596 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 00/0 : 3[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Channel 01/0 : 3[0] -> 2[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO Connected all trees +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO Connected all trees +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO Connected all trees +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888251:3888654 [3] NCCL INFO comm 0xb8e3fa0 rank 0 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2cba1ebe76552e9e - Init COMPLETE +ip-26-0-172-73:2574284:2574690 [2] NCCL INFO comm 0xbf9de60 rank 3 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3ca14ff2825f2b5f - Init COMPLETE +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO Connected all trees +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574283:2574688 [1] NCCL INFO comm 0xa6e6a00 rank 3 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e7aade2168ed94a - Init COMPLETE +ip-26-0-172-73:2574288:2574677 [6] NCCL INFO comm 0xa7745e0 rank 3 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2d469fc5421fa358 - Init COMPLETE +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO Connected all trees +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO Connected all trees +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713676:2714075 [3] NCCL INFO comm 0xa5b21f0 rank 2 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2cba1ebe76552e9e - Init COMPLETE +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO Connected all trees +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO Connected all trees +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574287:2574689 [5] NCCL INFO comm 0xbb58200 rank 3 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x32c4b1f092332bba - Init COMPLETE +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO Connected all trees +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114195:114583 [3] NCCL INFO Connected all trees +ip-26-0-172-142:114195:114583 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114195:114583 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-142:114195:114583 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574289:2574687 [7] NCCL INFO comm 0xba11fa0 rank 3 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x31dbc05cf868d21c - Init COMPLETE +ip-26-0-172-116:3888252:3888668 [4] NCCL INFO comm 0xbf3e4d0 rank 0 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x320345d6074aeb37 - Init COMPLETE +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO Connected all trees +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO Connected all trees +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888248:3888658 [0] NCCL INFO comm 0xc471eb0 rank 0 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x95b80b134b9b29b0 - Init COMPLETE +ip-26-0-172-142:114195:114583 [3] NCCL INFO comm 0xb7ebf10 rank 1 nranks 4 cudaDev 3 nvmlDev 3 busId 86000 commId 0x2cba1ebe76552e9e - Init COMPLETE +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO Connected all trees +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888249:3888662 [1] NCCL INFO comm 0xb2c5230 rank 0 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e7aade2168ed94a - Init COMPLETE +ip-26-0-172-73:2574286:2574691 [4] NCCL INFO comm 0xac20ea0 rank 3 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x320345d6074aeb37 - Init COMPLETE +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO Connected all trees +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888254:3888645 [6] NCCL INFO comm 0xc4a2c10 rank 0 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2d469fc5421fa358 - Init COMPLETE +ip-26-0-172-116:3888250:3888666 [2] NCCL INFO comm 0xb39e3f0 rank 0 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3ca14ff2825f2b5f - Init COMPLETE +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO Connected all trees +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574282:2574686 [0] NCCL INFO comm 0xc1ba6c0 rank 3 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x95b80b134b9b29b0 - Init COMPLETE +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO Connected all trees +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO Connected all trees +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114193:114598 [1] NCCL INFO Connected all trees +ip-26-0-172-142:114193:114598 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114193:114598 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-142:114193:114598 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO Connected all trees +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713675:2714080 [2] NCCL INFO comm 0xc2b5df0 rank 2 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3ca14ff2825f2b5f - Init COMPLETE +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO Connected all trees +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713674:2714078 [1] NCCL INFO comm 0xb885500 rank 2 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e7aade2168ed94a - Init COMPLETE +ip-26-0-172-142:114198:114582 [6] NCCL INFO Connected all trees +ip-26-0-172-142:114198:114582 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114198:114582 [6] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-142:114198:114582 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114193:114598 [1] NCCL INFO comm 0xb8cf2a0 rank 1 nranks 4 cudaDev 1 nvmlDev 1 busId 64000 commId 0x3e7aade2168ed94a - Init COMPLETE +ip-26-0-172-142:114194:114600 [2] NCCL INFO Connected all trees +ip-26-0-172-142:114194:114600 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114194:114600 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-142:114194:114600 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713679:2714066 [6] NCCL INFO comm 0xb3584a0 rank 2 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2d469fc5421fa358 - Init COMPLETE +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO Connected all trees +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888253:3888664 [5] NCCL INFO comm 0xb89f850 rank 0 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x32c4b1f092332bba - Init COMPLETE +ip-26-0-172-142:114198:114582 [6] NCCL INFO comm 0xbdd1ec0 rank 1 nranks 4 cudaDev 6 nvmlDev 6 busId b9000 commId 0x2d469fc5421fa358 - Init COMPLETE +ip-26-0-172-142:114194:114600 [2] NCCL INFO comm 0xb3cf0a0 rank 1 nranks 4 cudaDev 2 nvmlDev 2 busId 75000 commId 0x3ca14ff2825f2b5f - Init COMPLETE +ip-26-0-172-147:2713677:2714081 [4] NCCL INFO comm 0xb302cb0 rank 2 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x320345d6074aeb37 - Init COMPLETE +ip-26-0-172-142:114196:114601 [4] NCCL INFO Connected all trees +ip-26-0-172-142:114196:114601 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114196:114601 [4] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-142:114196:114601 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO Connected all trees +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO Connected all trees +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO Connected all trees +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713678:2714079 [5] NCCL INFO comm 0xbb10150 rank 2 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x32c4b1f092332bba - Init COMPLETE +ip-26-0-172-142:114196:114601 [4] NCCL INFO comm 0xa8b77a0 rank 1 nranks 4 cudaDev 4 nvmlDev 4 busId 97000 commId 0x320345d6074aeb37 - Init COMPLETE +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO Connected all trees +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888255:3888660 [7] NCCL INFO comm 0xb41e000 rank 0 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x31dbc05cf868d21c - Init COMPLETE +ip-26-0-172-147:2713680:2714077 [7] NCCL INFO comm 0xb489be0 rank 2 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x31dbc05cf868d21c - Init COMPLETE +ip-26-0-172-147:2713673:2714076 [0] NCCL INFO comm 0xbfc6260 rank 2 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x95b80b134b9b29b0 - Init COMPLETE +ip-26-0-172-142:114197:114599 [5] NCCL INFO Connected all trees +ip-26-0-172-142:114197:114599 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114197:114599 [5] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-142:114197:114599 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114199:114597 [7] NCCL INFO Connected all trees +ip-26-0-172-142:114199:114597 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114199:114597 [7] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-142:114199:114597 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114192:114596 [0] NCCL INFO Connected all trees +ip-26-0-172-142:114192:114596 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114192:114596 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 512 | 512 +ip-26-0-172-142:114192:114596 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114197:114599 [5] NCCL INFO comm 0xc247290 rank 1 nranks 4 cudaDev 5 nvmlDev 5 busId a8000 commId 0x32c4b1f092332bba - Init COMPLETE +ip-26-0-172-142:114199:114597 [7] NCCL INFO comm 0xb4538c0 rank 1 nranks 4 cudaDev 7 nvmlDev 7 busId ca000 commId 0x31dbc05cf868d21c - Init COMPLETE +ip-26-0-172-142:114192:114596 [0] NCCL INFO comm 0xb897c90 rank 1 nranks 4 cudaDev 0 nvmlDev 0 busId 53000 commId 0x95b80b134b9b29b0 - Init COMPLETE +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Total number of parameters: 8.86G (16896.51MiB) +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Local number of parameters: 2.73G (5200.14MiB) +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [After model building] Memory usage: 5200.16MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: No checkpoint path provided. +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Parametrizing model parameters using StandardParametrizator +12/28/2024 02:08:30 [INFO|DP=0|PP=3|TP=0|ip-26-0-172-73]: Local number of parameters: 2G (3808.10MiB) +12/28/2024 02:08:30 [INFO|DP=0|PP=2|TP=0|ip-26-0-172-147]: Local number of parameters: 1.95G (3712.12MiB) +12/28/2024 02:08:30 [INFO|DP=0|PP=3|TP=0|ip-26-0-172-73]: [After model building] Memory usage: 3808.12MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +12/28/2024 02:08:30 [INFO|DP=0|PP=2|TP=0|ip-26-0-172-147]: [After model building] Memory usage: 3712.14MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +12/28/2024 02:08:30 [INFO|DP=0|PP=1|TP=0|ip-26-0-172-142]: Local number of parameters: 2.19G (4176.14MiB) +12/28/2024 02:08:30 [INFO|DP=0|PP=1|TP=0|ip-26-0-172-142]: [After model building] Memory usage: 4176.16MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [Optimizer Building] Using LearningRateForSP as learning rate +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [ZeRO sharding] Size of optimizer params per rank: +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [ZeRO sharding] DP Rank 0 has 341M out of 2.73G (12.50%) params' optimizer states +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [ZeRO sharding] DP Rank 1 has 341M out of 2.73G (12.50%) params' optimizer states +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [ZeRO sharding] DP Rank 2 has 341M out of 2.73G (12.50%) params' optimizer states +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [ZeRO sharding] DP Rank 3 has 341M out of 2.73G (12.50%) params' optimizer states +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [ZeRO sharding] DP Rank 4 has 341M out of 2.73G (12.50%) params' optimizer states +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [ZeRO sharding] DP Rank 5 has 341M out of 2.73G (12.50%) params' optimizer states +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [ZeRO sharding] DP Rank 6 has 341M out of 2.73G (12.50%) params' optimizer states +12/28/2024 02:08:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [ZeRO sharding] DP Rank 7 has 341M out of 2.73G (12.50%) params' optimizer states +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.token_position_embeddings | PP: 0/4 | Block rank: 0 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.0 | PP: 0/4 | Block rank: 1 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.1 | PP: 0/4 | Block rank: 2 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.2 | PP: 0/4 | Block rank: 3 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.3 | PP: 0/4 | Block rank: 4 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.4 | PP: 0/4 | Block rank: 5 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.5 | PP: 0/4 | Block rank: 6 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.6 | PP: 0/4 | Block rank: 7 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.7 | PP: 0/4 | Block rank: 8 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.8 | PP: 0/4 | Block rank: 9 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.9 | PP: 1/4 | Block rank: 0 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.10 | PP: 1/4 | Block rank: 1 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.11 | PP: 1/4 | Block rank: 2 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.12 | PP: 1/4 | Block rank: 3 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.13 | PP: 1/4 | Block rank: 4 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.14 | PP: 1/4 | Block rank: 5 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.15 | PP: 1/4 | Block rank: 6 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.16 | PP: 1/4 | Block rank: 7 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.17 | PP: 1/4 | Block rank: 8 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.18 | PP: 2/4 | Block rank: 0 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.19 | PP: 2/4 | Block rank: 1 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.20 | PP: 2/4 | Block rank: 2 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.21 | PP: 2/4 | Block rank: 3 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.22 | PP: 2/4 | Block rank: 4 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.23 | PP: 2/4 | Block rank: 5 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.24 | PP: 2/4 | Block rank: 6 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.25 | PP: 2/4 | Block rank: 7 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.26 | PP: 3/4 | Block rank: 0 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.27 | PP: 3/4 | Block rank: 1 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.28 | PP: 3/4 | Block rank: 2 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.29 | PP: 3/4 | Block rank: 3 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.30 | PP: 3/4 | Block rank: 4 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.decoder.31 | PP: 3/4 | Block rank: 5 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.final_layer_norm | PP: 3/4 | Block rank: 6 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.lm_head | PP: 3/4 | Block rank: 7 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: model.cast_to_fp32 | PP: 3/4 | Block rank: 8 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: module_name: loss | PP: 3/4 | Block rank: 9 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Using dummy data generator +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [Training Plan] There are 1 training stages +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [Stage Stable Training Stage] start from step 1 +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: [Start training] datetime: 2024-12-28 02:08:31.099104 | mbs: 1 | grad_accum: 32 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/28/2024 02:08:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Memory usage: 16900.48MiB. Peak allocated 16900.48MiB. Peak reserved: 26362.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-142:114196:114623 [4] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Using network Libfabric +ip-26-0-172-142:114196:114623 [4] NCCL INFO comm 0xadac960 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2b8c87fc8eac6628 - Init START +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO comm 0x19176aa0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2b8c87fc8eac6628 - Init START +ip-26-0-172-142:114196:114623 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Using network Libfabric +ip-26-0-172-142:114199:114624 [7] NCCL INFO Using network Libfabric +ip-26-0-172-142:114199:114624 [7] NCCL INFO comm 0xbb22ed0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x390e51734d47d870 - Init START +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO comm 0x18860670 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x390e51734d47d870 - Init START +ip-26-0-172-142:114199:114624 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Using network Libfabric +ip-26-0-172-142:114192:114625 [0] NCCL INFO Using network Libfabric +ip-26-0-172-142:114192:114625 [0] NCCL INFO comm 0xbdc9ec0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2646922949db90cb - Init START +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO comm 0x1922f570 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2646922949db90cb - Init START +ip-26-0-172-142:114192:114625 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/253 +ip-26-0-172-142:114196:114623 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Using network Libfabric +ip-26-0-172-142:114195:114626 [3] NCCL INFO Using network Libfabric +ip-26-0-172-142:114195:114626 [3] NCCL INFO comm 0xbeb2720 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x393a54811051e3f1 - Init START +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO comm 0x1a715350 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x393a54811051e3f1 - Init START +ip-26-0-172-142:114195:114626 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Using network Libfabric +ip-26-0-172-142:114198:114627 [6] NCCL INFO Using network Libfabric +ip-26-0-172-142:114198:114627 [6] NCCL INFO comm 0xc310200 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7f0b237ed27928b2 - Init START +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO comm 0x19af1d10 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7f0b237ed27928b2 - Init START +ip-26-0-172-142:114198:114627 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114199:114624 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114192:114625 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-142:114196:114623 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-142:114196:114623 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Using network Libfabric +ip-26-0-172-142:114194:114630 [2] NCCL INFO Using network Libfabric +ip-26-0-172-142:114194:114630 [2] NCCL INFO comm 0xb90d130 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xaa287e63873b8e80 - Init START +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO comm 0x183fb5a0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xaa287e63873b8e80 - Init START +ip-26-0-172-142:114194:114630 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114198:114627 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114196:114623 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114623 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114623 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114623 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114623 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114623 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114623 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114623 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-142:114192:114625 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-142:114192:114625 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114195:114626 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-142:114199:114624 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-142:114199:114624 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-142:114193:114635 [1] NCCL INFO Using network Libfabric +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Using network Libfabric +ip-26-0-172-142:114193:114635 [1] NCCL INFO comm 0xbe08770 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1d2af5de9e855184 - Init START +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO comm 0x182cf830 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1d2af5de9e855184 - Init START +ip-26-0-172-142:114193:114635 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Using network Libfabric +ip-26-0-172-142:114197:114636 [5] NCCL INFO Using network Libfabric +ip-26-0-172-142:114197:114636 [5] NCCL INFO comm 0xc739cf0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe7dcf5ca6e59f7b8 - Init START +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO comm 0x1877f6e0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe7dcf5ca6e59f7b8 - Init START +ip-26-0-172-142:114197:114636 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114192:114625 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114625 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114625 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114625 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114625 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114625 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114625 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114625 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114199:114624 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114624 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114624 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114624 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114624 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114624 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114624 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114624 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114198:114627 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-142:114198:114627 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114194:114630 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114195:114626 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-142:114195:114626 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114198:114627 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114627 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114627 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114627 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114627 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114627 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114627 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114627 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114193:114635 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-142:114197:114636 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114195:114626 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114626 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114626 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114626 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114626 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114626 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114626 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114626 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-142:114194:114630 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-142:114194:114630 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114193:114635 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-142:114193:114635 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114197:114636 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-142:114197:114636 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114194:114630 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114630 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114630 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114630 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114630 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114630 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114630 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114630 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114193:114635 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114635 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114635 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114635 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114635 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114635 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114635 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114635 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114196:114623 [4] NCCL INFO Connected all rings +ip-26-0-172-142:114196:114623 [4] NCCL INFO Connected all trees +ip-26-0-172-142:114196:114623 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114196:114623 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114196:114623 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114197:114636 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114636 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114636 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114636 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114636 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114636 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114636 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114636 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114624 [7] NCCL INFO Connected all rings +ip-26-0-172-142:114199:114624 [7] NCCL INFO Connected all trees +ip-26-0-172-142:114199:114624 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114199:114624 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114199:114624 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114196:114623 [4] NCCL INFO comm 0xadac960 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2b8c87fc8eac6628 - Init COMPLETE +ip-26-0-172-142:114196:114648 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-142:114196:114648 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Connected all rings +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO Connected all trees +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114198:114627 [6] NCCL INFO Connected all rings +ip-26-0-172-142:114198:114627 [6] NCCL INFO Connected all trees +ip-26-0-172-142:114198:114627 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114198:114627 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114198:114627 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114192:114625 [0] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114625 [0] NCCL INFO Connected all trees +ip-26-0-172-142:114192:114625 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114192:114625 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114192:114625 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Connected all rings +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO Connected all trees +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114199:114624 [7] NCCL INFO comm 0xbb22ed0 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x390e51734d47d870 - Init COMPLETE +ip-26-0-172-142:114199:114649 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-142:114199:114649 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Connected all rings +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO Connected all trees +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888252:3888706 [4] NCCL INFO comm 0x19176aa0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0x2b8c87fc8eac6628 - Init COMPLETE +ip-26-0-172-116:3888252:3888739 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-116:3888252:3888739 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-142:114198:114627 [6] NCCL INFO comm 0xc310200 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7f0b237ed27928b2 - Init COMPLETE +ip-26-0-172-142:114198:114650 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-142:114198:114650 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Connected all rings +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO Connected all trees +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114192:114625 [0] NCCL INFO comm 0xbdc9ec0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2646922949db90cb - Init COMPLETE +ip-26-0-172-142:114192:114651 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-142:114192:114651 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-116:3888248:3888710 [0] NCCL INFO comm 0x1922f570 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0x2646922949db90cb - Init COMPLETE +ip-26-0-172-116:3888255:3888708 [7] NCCL INFO comm 0x18860670 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x390e51734d47d870 - Init COMPLETE +ip-26-0-172-116:3888255:3888741 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-116:3888248:3888740 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-116:3888255:3888741 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-116:3888248:3888740 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-116:3888254:3888716 [6] NCCL INFO comm 0x19af1d10 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x7f0b237ed27928b2 - Init COMPLETE +ip-26-0-172-116:3888254:3888742 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-116:3888254:3888742 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Connected all rings +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO Connected all trees +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114195:114626 [3] NCCL INFO Connected all rings +ip-26-0-172-142:114195:114626 [3] NCCL INFO Connected all trees +ip-26-0-172-142:114195:114626 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114195:114626 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114195:114626 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888251:3888713 [3] NCCL INFO comm 0x1a715350 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x393a54811051e3f1 - Init COMPLETE +ip-26-0-172-116:3888251:3888743 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-116:3888251:3888743 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-142:114195:114626 [3] NCCL INFO comm 0xbeb2720 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x393a54811051e3f1 - Init COMPLETE +ip-26-0-172-142:114195:114652 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-142:114195:114652 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-142:114194:114630 [2] NCCL INFO Connected all rings +ip-26-0-172-142:114194:114630 [2] NCCL INFO Connected all trees +ip-26-0-172-142:114194:114630 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114194:114630 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114194:114630 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Connected all rings +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO Connected all trees +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114193:114635 [1] NCCL INFO Connected all rings +ip-26-0-172-142:114193:114635 [1] NCCL INFO Connected all trees +ip-26-0-172-142:114193:114635 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114193:114635 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114193:114635 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114194:114630 [2] NCCL INFO comm 0xb90d130 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xaa287e63873b8e80 - Init COMPLETE +ip-26-0-172-142:114194:114653 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-142:114194:114653 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-142:114197:114636 [5] NCCL INFO Connected all rings +ip-26-0-172-142:114197:114636 [5] NCCL INFO Connected all trees +ip-26-0-172-142:114197:114636 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114197:114636 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114197:114636 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888250:3888720 [2] NCCL INFO comm 0x183fb5a0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0xaa287e63873b8e80 - Init COMPLETE +ip-26-0-172-116:3888250:3888744 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-116:3888250:3888744 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-142:114193:114635 [1] NCCL INFO comm 0xbe08770 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1d2af5de9e855184 - Init COMPLETE +ip-26-0-172-142:114193:114654 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-142:114193:114654 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Connected all rings +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO Connected all trees +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Connected all rings +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO Connected all trees +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114197:114636 [5] NCCL INFO comm 0xc739cf0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe7dcf5ca6e59f7b8 - Init COMPLETE +ip-26-0-172-142:114197:114655 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-142:114197:114655 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-116:3888249:3888726 [1] NCCL INFO comm 0x182cf830 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x1d2af5de9e855184 - Init COMPLETE +ip-26-0-172-116:3888249:3888745 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-116:3888249:3888745 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-116:3888253:3888728 [5] NCCL INFO comm 0x1877f6e0 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xe7dcf5ca6e59f7b8 - Init COMPLETE +ip-26-0-172-116:3888253:3888746 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-116:3888253:3888746 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-142:114196:114675 [4] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO comm 0xb72fca0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xdaeb54634c6fa68d - Init START +ip-26-0-172-142:114196:114675 [4] NCCL INFO comm 0x138ea490 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xdaeb54634c6fa68d - Init START +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114196:114675 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-142:114198:114677 [6] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO comm 0xba0e3e0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd22f43c548ad8b11 - Init START +ip-26-0-172-142:114198:114677 [6] NCCL INFO comm 0x147fcc10 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd22f43c548ad8b11 - Init START +ip-26-0-172-142:114198:114677 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-142:114199:114679 [7] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO comm 0xb8b3980 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x37bd69601a7d388f - Init START +ip-26-0-172-142:114199:114679 [7] NCCL INFO comm 0x141b2a20 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x37bd69601a7d388f - Init START +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114199:114679 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-142:114192:114681 [0] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO comm 0xc67a3f0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xab785a6072c784b8 - Init START +ip-26-0-172-142:114192:114681 [0] NCCL INFO comm 0x143f2530 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xab785a6072c784b8 - Init START +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114192:114681 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114199:114679 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114196:114675 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114196:114675 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-142:114196:114675 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-142:114199:114679 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-142:114199:114679 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-142:114198:114677 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-142:114195:114688 [3] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Using network Libfabric +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114198:114677 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-142:114198:114677 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO comm 0xac69ad0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x70a69eaac18bf32f - Init START +ip-26-0-172-142:114195:114688 [3] NCCL INFO comm 0x14674d20 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x70a69eaac18bf32f - Init START +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114195:114688 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114192:114681 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114192:114681 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-142:114192:114681 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114196:114675 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114199:114679 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-142:114198:114677 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114192:114681 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-142:114195:114688 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-142:114195:114688 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-142:114195:114688 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-142:114194:114696 [2] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Using network Libfabric +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO comm 0xc969a80 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x77c23b8f9e3acea8 - Init START +ip-26-0-172-142:114194:114696 [2] NCCL INFO comm 0x1412ad10 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x77c23b8f9e3acea8 - Init START +ip-26-0-172-142:114194:114696 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114193:114698 [1] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO comm 0xbf3adc0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2cb4117e87448fd1 - Init START +ip-26-0-172-142:114193:114698 [1] NCCL INFO comm 0x14b0bcd0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2cb4117e87448fd1 - Init START +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114193:114698 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-142:114197:114700 [5] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO comm 0xc1c5490 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa6387c16c9f079af - Init START +ip-26-0-172-142:114197:114700 [5] NCCL INFO comm 0x15334e60 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa6387c16c9f079af - Init START +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114197:114700 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-142:114195:114688 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-142:114193:114698 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-142:114193:114698 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-142:114193:114698 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114194:114696 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-142:114194:114696 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-142:114194:114696 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114197:114700 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-142:114197:114700 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-142:114197:114700 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114193:114698 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114194:114696 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114197:114700 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-142:114199:114679 [7] NCCL INFO Connected all rings +ip-26-0-172-142:114199:114679 [7] NCCL INFO Connected all trees +ip-26-0-172-142:114199:114679 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114199:114679 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114199:114679 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Connected all rings +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO Connected all trees +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Connected all rings +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO Connected all trees +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114198:114677 [6] NCCL INFO Connected all rings +ip-26-0-172-142:114198:114677 [6] NCCL INFO Connected all trees +ip-26-0-172-142:114198:114677 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114198:114677 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114198:114677 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Connected all rings +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO Connected all trees +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114196:114675 [4] NCCL INFO Connected all rings +ip-26-0-172-142:114196:114675 [4] NCCL INFO Connected all trees +ip-26-0-172-142:114196:114675 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114196:114675 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114196:114675 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114192:114681 [0] NCCL INFO Connected all rings +ip-26-0-172-142:114192:114681 [0] NCCL INFO Connected all trees +ip-26-0-172-142:114192:114681 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114192:114681 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114192:114681 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Connected all rings +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO Connected all trees +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713680:2714108 [7] NCCL INFO comm 0xb8b3980 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x37bd69601a7d388f - Init COMPLETE +ip-26-0-172-147:2713680:2714132 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-147:2713680:2714132 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-142:114199:114679 [7] NCCL INFO comm 0x141b2a20 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x37bd69601a7d388f - Init COMPLETE +ip-26-0-172-142:114199:114707 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-147:2713679:2714107 [6] NCCL INFO comm 0xba0e3e0 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd22f43c548ad8b11 - Init COMPLETE +ip-26-0-172-142:114199:114707 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-142:114198:114677 [6] NCCL INFO comm 0x147fcc10 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0xd22f43c548ad8b11 - Init COMPLETE +ip-26-0-172-147:2713679:2714133 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-147:2713679:2714133 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-142:114198:114708 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-142:114198:114708 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-147:2713677:2714106 [4] NCCL INFO comm 0xb72fca0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xdaeb54634c6fa68d - Init COMPLETE +ip-26-0-172-147:2713677:2714134 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-147:2713677:2714134 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-142:114196:114675 [4] NCCL INFO comm 0x138ea490 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xdaeb54634c6fa68d - Init COMPLETE +ip-26-0-172-142:114196:114709 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-142:114196:114709 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Connected all rings +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO Connected all trees +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114192:114681 [0] NCCL INFO comm 0x143f2530 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xab785a6072c784b8 - Init COMPLETE +ip-26-0-172-142:114192:114710 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-142:114192:114710 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-147:2713673:2714109 [0] NCCL INFO comm 0xc67a3f0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xab785a6072c784b8 - Init COMPLETE +ip-26-0-172-142:114195:114688 [3] NCCL INFO Connected all rings +ip-26-0-172-142:114195:114688 [3] NCCL INFO Connected all trees +ip-26-0-172-142:114195:114688 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114195:114688 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114195:114688 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713673:2714135 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-147:2713673:2714135 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-147:2713676:2714116 [3] NCCL INFO comm 0xac69ad0 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x70a69eaac18bf32f - Init COMPLETE +ip-26-0-172-147:2713676:2714136 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-147:2713676:2714136 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-142:114195:114688 [3] NCCL INFO comm 0x14674d20 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0x70a69eaac18bf32f - Init COMPLETE +ip-26-0-172-142:114195:114711 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-142:114195:114711 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-142:114193:114698 [1] NCCL INFO Connected all rings +ip-26-0-172-142:114193:114698 [1] NCCL INFO Connected all trees +ip-26-0-172-142:114193:114698 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114193:114698 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114193:114698 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114193:114698 [1] NCCL INFO comm 0x14b0bcd0 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2cb4117e87448fd1 - Init COMPLETE +ip-26-0-172-142:114193:114712 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-142:114193:114712 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Connected all rings +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO Connected all trees +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114194:114696 [2] NCCL INFO Connected all rings +ip-26-0-172-142:114194:114696 [2] NCCL INFO Connected all trees +ip-26-0-172-142:114194:114696 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114194:114696 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114194:114696 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713674:2714124 [1] NCCL INFO comm 0xbf3adc0 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0x2cb4117e87448fd1 - Init COMPLETE +ip-26-0-172-142:114194:114696 [2] NCCL INFO comm 0x1412ad10 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x77c23b8f9e3acea8 - Init COMPLETE +ip-26-0-172-142:114194:114713 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-142:114194:114713 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-147:2713674:2714137 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-147:2713674:2714137 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Connected all rings +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO Connected all trees +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713675:2714123 [2] NCCL INFO comm 0xc969a80 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x77c23b8f9e3acea8 - Init COMPLETE +ip-26-0-172-147:2713675:2714138 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-147:2713675:2714138 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-142:114197:114700 [5] NCCL INFO Connected all rings +ip-26-0-172-142:114197:114700 [5] NCCL INFO Connected all trees +ip-26-0-172-142:114197:114700 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-142:114197:114700 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-142:114197:114700 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Connected all rings +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO Connected all trees +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-142:114197:114700 [5] NCCL INFO comm 0x15334e60 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa6387c16c9f079af - Init COMPLETE +ip-26-0-172-142:114197:114714 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-142:114197:114714 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-147:2713678:2714125 [5] NCCL INFO comm 0xc1c5490 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0xa6387c16c9f079af - Init COMPLETE +ip-26-0-172-147:2713678:2714139 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-147:2713678:2714139 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Using network Libfabric +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO comm 0xb3b9a90 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x75d91a81b173980c - Init START +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO comm 0x14297930 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x75d91a81b173980c - Init START +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO comm 0xb238510 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf83b32094b25ad83 - Init START +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO comm 0x130eac20 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf83b32094b25ad83 - Init START +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO comm 0xb864ad0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb4b4305d2ecc2662 - Init START +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO comm 0x13d18ee0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb4b4305d2ecc2662 - Init START +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO comm 0xc654e70 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5afd71dc383e560d - Init START +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO comm 0x13fafa70 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5afd71dc383e560d - Init START +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO comm 0xce1c9a0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcb03e7b0f5eb8373 - Init START +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO comm 0x14e82570 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcb03e7b0f5eb8373 - Init START +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Setting affinity for GPU 4 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO P2P Chunksize set to 131072 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO comm 0xcbe2b40 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2276036be0f6096f - Init START +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO comm 0x14fb49f0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2276036be0f6096f - Init START +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Setting affinity for GPU 3 to ffff,ffffffff +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO comm 0xb32c200 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf67a9fe2d7a70bb3 - Init START +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO comm 0x14753a30 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf67a9fe2d7a70bb3 - Init START +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Using network Libfabric +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Using network Libfabric +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO comm 0xc79bdc0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8c2bac5ef99fb68e - Init START +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO comm 0x1484a830 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8c2bac5ef99fb68e - Init START +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO NCCL_TOPO_FILE set by environment to /proc/self/fd/216 +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Setting affinity for GPU 6 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Setting affinity for GPU 7 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Setting affinity for GPU 0 to ffff,ffffffff +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 00/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Channel 00/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Channel 01/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 01/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Channel 02/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 02/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Channel 03/0 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Channel 03/0 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 00/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 01/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 02/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 03/0 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 00/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 01/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 02/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Channel 03/0 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 00/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 01/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 02/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 03/0 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 00/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 01/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 02/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Channel 03/0 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 00/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 01/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 02/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 03/0 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 00/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 01/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 02/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Channel 03/0 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 00/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 01/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 02/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 03/0 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 00/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 01/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 02/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Channel 03/0 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Setting affinity for GPU 2 to ffff,ffffffff +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Setting affinity for GPU 1 to ffff,ffffffff +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Setting affinity for GPU 5 to ffffffff,ffff0000,00000000 +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Trees [0] -1/-1/-1->1->0 [1] -1/-1/-1->1->0 [2] 0/-1/-1->1->-1 [3] 0/-1/-1->1->-1 +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 00/04 : 0 1 +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 01/04 : 0 1 +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 02/04 : 0 1 +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 03/04 : 0 1 +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1 [2] -1/-1/-1->0->1 [3] -1/-1/-1->0->1 +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO P2P Chunksize set to 131072 +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 00/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 01/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 02/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 03/0 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 00/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 01/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 02/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Channel 03/0 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 00/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 01/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 02/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 03/0 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 00/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 01/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 02/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Channel 03/0 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 00/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 01/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 02/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 03/0 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 00/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 01/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 02/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Channel 03/0 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Connected all rings +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO Connected all trees +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Connected all rings +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO Connected all trees +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Connected all rings +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO Connected all trees +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Connected all rings +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO Connected all trees +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Connected all rings +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO Connected all trees +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Connected all rings +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO Connected all trees +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Connected all rings +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO Connected all trees +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Connected all rings +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO Connected all trees +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713676:2714164 [3] NCCL INFO comm 0x130eac20 rank 0 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf83b32094b25ad83 - Init COMPLETE +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Connected all rings +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO Connected all trees +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713676:2714192 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-147:2713676:2714192 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-73:2574288:2574730 [6] NCCL INFO comm 0xb3b9a90 rank 1 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x75d91a81b173980c - Init COMPLETE +ip-26-0-172-73:2574288:2574756 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-73:2574288:2574756 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-147:2713680:2714166 [7] NCCL INFO comm 0x13fafa70 rank 0 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5afd71dc383e560d - Init COMPLETE +ip-26-0-172-147:2713680:2714193 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-147:2713680:2714193 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-73:2574285:2574731 [3] NCCL INFO comm 0xb238510 rank 1 nranks 2 cudaDev 3 nvmlDev 3 busId 86000 commId 0xf83b32094b25ad83 - Init COMPLETE +ip-26-0-172-73:2574285:2574757 [3] NCCL INFO Channel 02/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-73:2574285:2574757 [3] NCCL INFO Channel 03/1 : 0[3] -> 1[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-73:2574286:2574732 [4] NCCL INFO comm 0xb864ad0 rank 1 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb4b4305d2ecc2662 - Init COMPLETE +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Connected all rings +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO Connected all trees +ip-26-0-172-73:2574286:2574758 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574286:2574758 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-73:2574282:2574734 [0] NCCL INFO comm 0xce1c9a0 rank 1 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcb03e7b0f5eb8373 - Init COMPLETE +ip-26-0-172-73:2574282:2574759 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-73:2574282:2574759 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-73:2574289:2574733 [7] NCCL INFO comm 0xc654e70 rank 1 nranks 2 cudaDev 7 nvmlDev 7 busId ca000 commId 0x5afd71dc383e560d - Init COMPLETE +ip-26-0-172-147:2713677:2714163 [4] NCCL INFO comm 0x13d18ee0 rank 0 nranks 2 cudaDev 4 nvmlDev 4 busId 97000 commId 0xb4b4305d2ecc2662 - Init COMPLETE +ip-26-0-172-73:2574289:2574760 [7] NCCL INFO Channel 02/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-73:2574289:2574760 [7] NCCL INFO Channel 03/1 : 0[7] -> 1[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-147:2713677:2714194 [4] NCCL INFO Channel 02/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-147:2713677:2714194 [4] NCCL INFO Channel 03/1 : 0[4] -> 1[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-147:2713679:2714160 [6] NCCL INFO comm 0x14297930 rank 0 nranks 2 cudaDev 6 nvmlDev 6 busId b9000 commId 0x75d91a81b173980c - Init COMPLETE +ip-26-0-172-147:2713679:2714195 [6] NCCL INFO Channel 02/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-147:2713679:2714195 [6] NCCL INFO Channel 03/1 : 0[6] -> 1[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-147:2713673:2714168 [0] NCCL INFO comm 0x14e82570 rank 0 nranks 2 cudaDev 0 nvmlDev 0 busId 53000 commId 0xcb03e7b0f5eb8373 - Init COMPLETE +ip-26-0-172-147:2713673:2714196 [0] NCCL INFO Channel 02/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-147:2713673:2714196 [0] NCCL INFO Channel 03/1 : 0[0] -> 1[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Connected all rings +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO Connected all trees +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574284:2574737 [2] NCCL INFO comm 0xcbe2b40 rank 1 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2276036be0f6096f - Init COMPLETE +ip-26-0-172-73:2574284:2574761 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-73:2574284:2574761 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Connected all rings +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO Connected all trees +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713675:2714172 [2] NCCL INFO comm 0x14fb49f0 rank 0 nranks 2 cudaDev 2 nvmlDev 2 busId 75000 commId 0x2276036be0f6096f - Init COMPLETE +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Connected all rings +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO Connected all trees +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713675:2714197 [2] NCCL INFO Channel 02/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-147:2713675:2714197 [2] NCCL INFO Channel 03/1 : 0[2] -> 1[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Connected all rings +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO Connected all trees +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-147:2713674:2714174 [1] NCCL INFO comm 0x14753a30 rank 0 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf67a9fe2d7a70bb3 - Init COMPLETE +ip-26-0-172-147:2713674:2714198 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-147:2713674:2714198 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-73:2574283:2574738 [1] NCCL INFO comm 0xb32c200 rank 1 nranks 2 cudaDev 1 nvmlDev 1 busId 64000 commId 0xf67a9fe2d7a70bb3 - Init COMPLETE +ip-26-0-172-73:2574283:2574762 [1] NCCL INFO Channel 02/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Connected all rings +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO Connected all trees +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574283:2574762 [1] NCCL INFO Channel 03/1 : 0[1] -> 1[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Connected all rings +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO Connected all trees +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO NCCL_PROTO set by environment to simple +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO threadThresholds 8/8/64 | 16/8/64 | 512 | 512 +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO 4 coll channels, 0 nvls channels, 4 p2p channels, 2 p2p channels per peer +ip-26-0-172-73:2574287:2574740 [5] NCCL INFO comm 0xc79bdc0 rank 1 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8c2bac5ef99fb68e - Init COMPLETE +ip-26-0-172-147:2713678:2714177 [5] NCCL INFO comm 0x1484a830 rank 0 nranks 2 cudaDev 5 nvmlDev 5 busId a8000 commId 0x8c2bac5ef99fb68e - Init COMPLETE +ip-26-0-172-73:2574287:2574763 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-73:2574287:2574763 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-147:2713678:2714199 [5] NCCL INFO Channel 02/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-147:2713678:2714199 [5] NCCL INFO Channel 03/1 : 0[5] -> 1[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-147:2713676:2714267 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-147:2713676:2714267 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-147:2713677:2714268 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-147:2713677:2714268 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-147:2713680:2714269 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-147:2713680:2714269 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-147:2713679:2714270 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-147:2713679:2714270 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-147:2713673:2714271 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-147:2713673:2714271 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-73:2574285:2574847 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-73:2574285:2574847 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-73:2574282:2574848 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-73:2574282:2574848 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-73:2574288:2574849 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-73:2574289:2574850 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-73:2574288:2574849 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-73:2574289:2574850 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-73:2574286:2574851 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-73:2574286:2574851 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-147:2713675:2714273 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-147:2713675:2714273 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-147:2713674:2714274 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-147:2713674:2714274 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-147:2713678:2714275 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-147:2713678:2714275 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-73:2574284:2574853 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-73:2574284:2574853 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-73:2574283:2574854 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-73:2574283:2574854 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +ip-26-0-172-73:2574287:2574855 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-73:2574287:2574855 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-142:114198:114787 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-142:114198:114787 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-142:114196:114789 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-142:114196:114789 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-142:114195:114790 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-142:114195:114790 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-142:114199:114791 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-142:114199:114791 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-142:114192:114792 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-142:114192:114792 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-147:2713679:2714278 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-147:2713679:2714278 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-147:2713677:2714279 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-147:2713677:2714279 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-147:2713676:2714280 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-147:2713676:2714280 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-147:2713680:2714281 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-147:2713680:2714281 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-147:2713673:2714282 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-147:2713673:2714282 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-142:114194:114794 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-142:114194:114794 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-142:114197:114795 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-142:114197:114795 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-142:114193:114796 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-142:114193:114796 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-147:2713675:2714283 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-147:2713675:2714283 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-147:2713678:2714284 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-147:2713678:2714284 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-147:2713674:2714285 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-147:2713674:2714285 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-116:3888254:3888825 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-116:3888254:3888825 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [receive] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-142:114198:114798 [6] NCCL INFO Channel 02/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-142:114198:114798 [6] NCCL INFO Channel 03/1 : 1[6] -> 0[6] [send] via NET/Libfabric/6/GDRDMA/Shared +ip-26-0-172-116:3888251:3888826 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-116:3888251:3888826 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [receive] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-116:3888252:3888827 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-116:3888252:3888827 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [receive] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-116:3888255:3888828 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-116:3888255:3888828 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [receive] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-116:3888248:3888829 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-116:3888248:3888829 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [receive] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-142:114196:114800 [4] NCCL INFO Channel 02/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-142:114196:114800 [4] NCCL INFO Channel 03/1 : 1[4] -> 0[4] [send] via NET/Libfabric/4/GDRDMA/Shared +ip-26-0-172-142:114195:114801 [3] NCCL INFO Channel 02/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-142:114195:114801 [3] NCCL INFO Channel 03/1 : 1[3] -> 0[3] [send] via NET/Libfabric/3/GDRDMA/Shared +ip-26-0-172-142:114199:114802 [7] NCCL INFO Channel 02/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-142:114199:114802 [7] NCCL INFO Channel 03/1 : 1[7] -> 0[7] [send] via NET/Libfabric/7/GDRDMA/Shared +ip-26-0-172-142:114192:114803 [0] NCCL INFO Channel 02/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-142:114192:114803 [0] NCCL INFO Channel 03/1 : 1[0] -> 0[0] [send] via NET/Libfabric/0/GDRDMA/Shared +ip-26-0-172-116:3888253:3888831 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-116:3888253:3888831 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [receive] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-116:3888250:3888832 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-116:3888250:3888832 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [receive] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-116:3888249:3888833 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-116:3888249:3888833 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [receive] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-142:114197:114804 [5] NCCL INFO Channel 02/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-142:114197:114804 [5] NCCL INFO Channel 03/1 : 1[5] -> 0[5] [send] via NET/Libfabric/5/GDRDMA/Shared +ip-26-0-172-142:114193:114805 [1] NCCL INFO Channel 02/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-142:114193:114805 [1] NCCL INFO Channel 03/1 : 1[1] -> 0[1] [send] via NET/Libfabric/1/GDRDMA/Shared +ip-26-0-172-142:114194:114806 [2] NCCL INFO Channel 02/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +ip-26-0-172-142:114194:114806 [2] NCCL INFO Channel 03/1 : 1[2] -> 0[2] [send] via NET/Libfabric/2/GDRDMA/Shared +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +12/28/2024 02:09:16 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Memory usage: 16973.51MiB. Peak allocated 45008.79MiB. Peak reserved: 46598.00MiB +12/28/2024 02:09:17 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Memory usage: 19573.61MiB. Peak allocated 30623.91MiB. Peak reserved: 57016.00MiB +12/28/2024 02:09:17 [INFO|DP=0|PP=3|TP=0|ip-26-0-172-73]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 44.4K | tokens_per_sec: 23.6K | tokens_per_sec_per_gpu: 738 | global_batch_size: 256 | lm_loss: 12.6 | lr: 0.00015 | model_tflops_per_gpu: 41.6 | hardware_tflops_per_gpu: 41.6 | grad_norm: 3.9 | cuda_memory_allocated: 15G | cuda_max_memory_reserved: 29.5G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.6G | hd_free_memory_tb: 244G +12/28/2024 02:09:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Memory usage: 19573.61MiB. Peak allocated 47608.88MiB. Peak reserved: 57016.00MiB +12/28/2024 02:09:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Memory usage: 19573.61MiB. Peak allocated 30623.91MiB. Peak reserved: 57016.00MiB +12/28/2024 02:09:52 [INFO|DP=0|PP=3|TP=0|ip-26-0-172-73]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 35.5K | tokens_per_sec: 29.5K | tokens_per_sec_per_gpu: 922 | global_batch_size: 256 | lm_loss: 12.6 | lr: 0.0003 | model_tflops_per_gpu: 52 | hardware_tflops_per_gpu: 52 | grad_norm: 3.9 | cuda_memory_allocated: 15G | cuda_max_memory_reserved: 29.5G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.6G | hd_free_memory_tb: 244G +12/28/2024 02:10:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Memory usage: 19573.61MiB. Peak allocated 47608.88MiB. Peak reserved: 57016.00MiB +12/28/2024 02:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/28/2024 02:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: | -------- | -------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | ----------------------------------- | -------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/28/2024 02:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: | 13848550 | 8.86G_dp8_tp1_pp4_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k | 4 | 4096 | 1 | 32 | 256 | 51.73 | 51.73 | 917.68 | 160.78 | 102.21 | 99.35 | 422.78 | 241.74 | 241.60 | 29.91 | 55.68 | 8 | 4 | 1 | OneForwardOneBackwardPipelineEngine | REDUCE_SCATTER | True | 4096 | silu | 32 | 32 | 32 | 4096 | 131072 | False | torch.bfloat16 | 1 | 25 | True | 8.86G | 2.73G | +12/28/2024 02:10:28 [INFO|DP=0|PP=3|TP=0|ip-26-0-172-73]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 35.7K | tokens_per_sec: 29.4K | tokens_per_sec_per_gpu: 918 | global_batch_size: 256 | lm_loss: 12.6 | lr: 0.000296 | model_tflops_per_gpu: 51.7 | hardware_tflops_per_gpu: 51.7 | grad_norm: 3.87 | cuda_memory_allocated: 15G | cuda_max_memory_reserved: 29.5G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.6G | hd_free_memory_tb: 244G +12/28/2024 02:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final2.csv +12/28/2024 02:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Acquired lock for CSV file: benchmark/results/bench_final2.csv +12/28/2024 02:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Successfully wrote to CSV file: benchmark/results/bench_final2.csv. Releasing lock... +12/28/2024 02:10:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-116]: Throughput logging complete +ip-26-0-172-116:3888255:3888564 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-116:3888253:3888563 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-116:3888252:3888565 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-116:3888250:3888566 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-116:3888251:3888568 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-116:3888249:3888570 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-116:3888255:3888433 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-116:3888253:3888435 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-116:3888252:3888434 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-116:3888249:3888429 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-116:3888251:3888431 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-116:3888250:3888430 [0] NCCL INFO [Service thread] Connection closed by localRank 0 +ip-26-0-172-147:2713673:2714027 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-147:2713680:2713989 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-147:2713679:2713986 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-147:2713678:2713988 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-147:2713676:2713992 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-147:2713673:2713991 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-147:2713680:2713852 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-147:2713679:2713853 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-147:2713678:2713854 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-147:2713676:2713858 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-147:2713673:2713859 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-147:2713673:2714027 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-172-147:2713680:2713989 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-172-147:2713679:2713986 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-172-147:2713678:2713988 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-172-147:2713676:2713992 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-172-147:2713673:2713991 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-172-147:2713680:2713852 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-172-147:2713679:2713853 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-172-147:2713678:2713854 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-172-147:2713676:2713858 [0] NCCL INFO [Service thread] Connection closed by localRank 2 +ip-26-0-172-147:2713678:2713988 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-172-147:2713678:2713854 [0] NCCL INFO [Service thread] Connection closed by localRank 4 +ip-26-0-172-142:114192:114543 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114196:114508 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114197:114511 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114195:114509 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114194:114510 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114198:114504 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114192:114506 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114197:114374 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114196:114373 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114195:114375 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114194:114377 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114198:114372 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114192:114370 [0] NCCL INFO [Service thread] Connection closed by localRank 7 +ip-26-0-172-142:114192:114543 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-142:114197:114511 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-142:114196:114508 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-142:114194:114510 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-142:114192:114506 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-142:114196:114373 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-142:114197:114374 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-142:114194:114377 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +ip-26-0-172-142:114192:114370 [0] NCCL INFO [Service thread] Connection closed by localRank 1 +[2024-12-28 02:10:48,526] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2574283 closing signal SIGTERM +[2024-12-28 02:10:48,526] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2574284 closing signal SIGTERM +[2024-12-28 02:10:48,526] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2574285 closing signal SIGTERM +[2024-12-28 02:10:48,526] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2574286 closing signal SIGTERM +[2024-12-28 02:10:48,527] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888249 closing signal SIGTERM +[2024-12-28 02:10:48,526] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2574289 closing signal SIGTERM +[2024-12-28 02:10:48,527] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888250 closing signal SIGTERM +[2024-12-28 02:10:48,527] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888251 closing signal SIGTERM +[2024-12-28 02:10:48,527] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888252 closing signal SIGTERM +[2024-12-28 02:10:48,527] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888253 closing signal SIGTERM +[2024-12-28 02:10:48,527] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888254 closing signal SIGTERM +[2024-12-28 02:10:48,527] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888255 closing signal SIGTERM +[2024-12-28 02:10:50,593] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 2574282) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-28_02:10:48 + host : ip-26-0-172-73.ec2.internal + rank : 29 (local_rank: 5) + exitcode : 1 (pid: 2574287) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2024-12-28_02:10:48 + host : ip-26-0-172-73.ec2.internal + rank : 30 (local_rank: 6) + exitcode : 1 (pid: 2574288) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-28_02:10:48 + host : ip-26-0-172-73.ec2.internal + rank : 24 (local_rank: 0) + exitcode : 1 (pid: 2574282) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-172-73: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=13848550.0 +[2024-12-28 02:10:50,923] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-28 02:10:50,924] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114192 closing signal SIGTERM +[2024-12-28 02:10:50,924] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114193 closing signal SIGTERM +[2024-12-28 02:10:50,924] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114194 closing signal SIGTERM +[2024-12-28 02:10:50,924] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114195 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-28 02:10:50,924] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114196 closing signal SIGTERM +[2024-12-28 02:10:50,924] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114197 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-28 02:10:50,924] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114198 closing signal SIGTERM +[2024-12-28 02:10:50,924] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114199 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2713673 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2713674 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2713675 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2713676 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888249 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2713677 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2713678 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888251 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2713679 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2713680 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888252 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888253 closing signal SIGTERM +[2024-12-28 02:10:50,925] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3888255 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 3888176 got signal: 15 +srun: error: ip-26-0-172-116: task 1: Exited with exit code 1 +[2024-12-28 02:10:53,440] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-142.ec2.internal_114120_0' has failed to send a keep-alive heartbeat to the rendezvous '13848550' due to an error of type RendezvousConnectionError. +[2024-12-28 02:10:53,485] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-147.ec2.internal_2713602_0' has failed to send a keep-alive heartbeat to the rendezvous '13848550' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 114120 got signal: 15 +srun: error: ip-26-0-172-142: task 2: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2713602 got signal: 15 +srun: error: ip-26-0-172-147: task 3: Exited with exit code 1 diff --git a/logs/13849583-bench_1.34G_dp8_tp4_pp8_acc16_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13849583-bench_1.34G_dp8_tp4_pp8_acc16_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..9bde3c51c6ecfcfcfc6b0a550e642b496c8f04cf --- /dev/null +++ b/logs/13849583-bench_1.34G_dp8_tp4_pp8_acc16_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:30ff65e5dcda36b3ba60f3d45dcdfadf67ccaf2c845b57fdd83add48a8d5441b +size 16735842 diff --git a/logs/13849695-bench_80G_dp16_tp8_pp2_acc2_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13849695-bench_80G_dp16_tp8_pp2_acc2_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..d85a6ddb13b2d7de0f9c67fc6f75d6dec2ec71fb --- /dev/null +++ b/logs/13849695-bench_80G_dp16_tp8_pp2_acc2_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4c17023ac5e2d39083bf27caacd2a8e20d51d5c506f158e25eb588442c5c9f4c +size 11774731 diff --git a/logs/13849751-bench_80G_dp64_tp2_pp2_acc4_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13849751-bench_80G_dp64_tp2_pp2_acc4_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..0d04d89a156841e5049c4c4876951cb94870b604 --- /dev/null +++ b/logs/13849751-bench_80G_dp64_tp2_pp2_acc4_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c9600f7d6d55051c7ca21bd82888244438794324673e5649f468b22673bb5fd3 +size 14438974 diff --git a/logs/13849976-bench_469G_dp8_tp32_pp2_acc1_mbs32_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/13849976-bench_469G_dp8_tp32_pp2_acc1_mbs32_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..ab99d6fe25eafd68c60442efa1fa7fd39d7f2c81 --- /dev/null +++ b/logs/13849976-bench_469G_dp8_tp32_pp2_acc1_mbs32_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d2501bb38bc6d7068cbbc4c2912373d6c8fb3a9ff930637d16cbe49e2385f452 +size 29089166 diff --git a/logs/14016402-bench_8.86G_dp16_tp32_pp1_acc8_mbs2_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/14016402-bench_8.86G_dp16_tp32_pp1_acc8_mbs2_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..86f5f332ad0a800a8fdeb45918afbc893ed9c842 --- /dev/null +++ b/logs/14016402-bench_8.86G_dp16_tp32_pp1_acc8_mbs2_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,12238 @@ ++ '[' -z 14016402 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-163-[134,147,158,220,226,236],ip-26-0-164-[0,18,45,75,187,207,236],ip-26-0-165-[24,38,59,131,164,202,213],ip-26-0-166-15,ip-26-0-168-[30,34,52,95,120,238],ip-26-0-169-[86,132,139,207,239,247],ip-26-0-170-[31,132,143,160],ip-26-0-171-[21,56,62,88,102,168,230,249],ip-26-0-172-[57,73,116,142,147,252],ip-26-0-173-[7,121,202,246],ip-26-0-174-[36,100,186,196,240],ip-26-0-175-[19,34,132,165]' ++ export 'NODELIST=ip-26-0-163-134 +ip-26-0-163-147 +ip-26-0-163-158 +ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165' ++ NODELIST='ip-26-0-163-134 +ip-26-0-163-147 +ip-26-0-163-158 +ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-163-[134,147,158,220,226,236],ip-26-0-164-[0,18,45,75,187,207,236],ip-26-0-165-[24,38,59,131,164,202,213],ip-26-0-166-15,ip-26-0-168-[30,34,52,95,120,238],ip-26-0-169-[86,132,139,207,239,247],ip-26-0-170-[31,132,143,160],ip-26-0-171-[21,56,62,88,102,168,230,249],ip-26-0-172-[57,73,116,142,147,252],ip-26-0-173-[7,121,202,246],ip-26-0-174-[36,100,186,196,240],ip-26-0-175-[19,34,132,165]' ++ export MASTER_NODE=ip-26-0-163-134 ++ MASTER_NODE=ip-26-0-163-134 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=64 ++ NNODES=64 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=512 ++ WORLD_SIZE=512 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-163-134' +Master node: ip-26-0-163-134 ++ echo 'All nodes: ip-26-0-163-134 +ip-26-0-163-147 +ip-26-0-163-158 +ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165' +All nodes: ip-26-0-163-134 +ip-26-0-163-147 +ip-26-0-163-158 +ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165 ++ echo 'World size: 512' +World size: 512 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=64 --nproc_per_node=8 --rdzv_id=14016402 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-163-134:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_8.86G_dp16_tp32_pp1_acc8_mbs2_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,821] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,824] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,824] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,824] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,824] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,825] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,823] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,825] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,825] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,825] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,825] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,824] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,824] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,824] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,824] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,826] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,827] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,828] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,828] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,828] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,828] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,828] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,829] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,829] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,829] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,829] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,829] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,829] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,829] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,829] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,829] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,829] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,830] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,831] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,831] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,831] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,831] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,830] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,830] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,830] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,830] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,830] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,837] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,837] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,837] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,837] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,837] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,840] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,840] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,840] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,840] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,840] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,839] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,839] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,839] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,839] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,839] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,843] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,843] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,843] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,843] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,843] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,845] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,845] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,845] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,845] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,845] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,846] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,846] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,846] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,846] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,846] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,846] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,846] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,846] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,846] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,846] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,848] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,848] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,848] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,848] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,848] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,848] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,848] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,848] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,848] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,848] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,850] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,850] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,850] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,850] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,850] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,853] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,853] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,853] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,853] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,853] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,854] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,854] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,854] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,854] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,854] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,855] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,855] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,855] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,855] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,855] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,858] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,858] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,858] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,858] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,858] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,862] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,862] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,862] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,862] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,862] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,873] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,874] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,874] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,874] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,874] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,874] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,877] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,877] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,877] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,877] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,877] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,877] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,879] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,879] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,879] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,879] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,879] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,882] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,883] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,883] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,883] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,883] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,884] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,884] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,884] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,884] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,884] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,887] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,887] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,887] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,887] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,887] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,891] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,891] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,891] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,891] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,891] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,892] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,892] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,892] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,892] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,892] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,895] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,895] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,895] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,895] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,895] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,904] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,905] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,905] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,905] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,905] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,905] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,905] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,905] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,905] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,905] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,873] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,873] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,873] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,873] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,916] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,916] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,916] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,916] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,916] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,877] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,877] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,877] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,877] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,924] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,924] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,924] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,924] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,924] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,928] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,928] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,928] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,928] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,928] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,931] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,931] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,931] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,931] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,931] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,944] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,944] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,944] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,944] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,944] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,946] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,946] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,946] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,946] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,946] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,946] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,946] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,946] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,946] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,946] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,956] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,957] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,957] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,957] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,957] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,960] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,960] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,960] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,960] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,960] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,970] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,970] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,970] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,970] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,970] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,970] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,996] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,996] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,996] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,996] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,996] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:56,000] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:56,001] torch.distributed.run: [WARNING] +[2024-12-30 17:47:56,001] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:56,001] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:56,001] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:56,008] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:55,970] torch.distributed.run: [WARNING] +[2024-12-30 17:47:55,970] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:55,970] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:55,970] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:56,015] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:56,016] torch.distributed.run: [WARNING] +[2024-12-30 17:47:56,016] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:56,016] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:56,016] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:56,008] torch.distributed.run: [WARNING] +[2024-12-30 17:47:56,008] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:56,008] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:56,008] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:59,519] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-30 17:47:59,519] torch.distributed.run: [WARNING] +[2024-12-30 17:47:59,519] torch.distributed.run: [WARNING] ***************************************** +[2024-12-30 17:47:59,519] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-30 17:47:59,519] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Config: +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Config(general=GeneralArgs(project='debug', +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: run='8.86G_dp16_tp32_pp1_acc8_mbs2_seq4096_zero0_tpmodeRED_vocab131k', +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: seed=42, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: step=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: consumed_train_samples=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: ignore_sanity_checks=True), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: parallelism=ParallelismArgs(dp=16, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pp=1, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tp=32, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pp_engine=, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tp_mode=, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tp_linear_async_communication=True, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: recompute_layer=False, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tp_recompute_allgather=True, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: expert_parallel_size=1), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: eos_token_id=0, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: hidden_act='silu', +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: hidden_size=4096, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: initializer_range=0.02, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: intermediate_size=14336, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: is_llama_config=True, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: max_position_embeddings=4096, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_attention_heads=32, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_hidden_layers=32, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_key_value_heads=32, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pad_token_id=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pretraining_tp=1, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rms_norm_eps=1e-05, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_scaling=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_theta=10000.0, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_interleaved=False, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tie_word_embeddings=False, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: use_cache=True, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: vocab_size=131072), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: init_method=RandomInit(std=0.02), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: dtype=torch.bfloat16, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: make_vocab_size_divisible_by=1, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: ddp_bucket_cap_mb=25), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tokenizer_revision=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tokenizer_max_length=None), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: checkpoint_interval=10000, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: save_initial_state=False, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: save_final_state=False, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: resume_checkpoint_path=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: checkpoints_path_is_shared_file_system=False), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: logging=LoggingArgs(log_level='info', +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: log_level_replica='info', +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: iteration_step_info_interval=1), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tokens=TokensArgs(sequence_length=4096, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: train_steps=100, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: micro_batch_size=2, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: batch_accumulation_per_replica=8, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: val_check_interval=100, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: limit_val_batches=0, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: limit_test_batches=0), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: adam_beta1=0.9, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: adam_beta2=0.95, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: torch_adam_is_fused=True, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: name='adamW'), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: zero_stage=0, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: weight_decay=0.01, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: clip_grad=1.0, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: accumulate_grad_in_fp32=True, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lr_warmup_steps=2, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lr_warmup_style='linear', +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lr_decay_style='cosine', +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lr_decay_steps=13, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lr_decay_starting_step=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: min_decay_lr=1e-05)), +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: start_training_step=1, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: data=DataArgs(dataset=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: seed=42, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_loading_workers=1))], +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: profiler=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lighteval=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: s3_upload=None) +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Model Config: +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: LlamaConfig(bos_token_id=0, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: eos_token_id=0, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: hidden_act='silu', +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: hidden_size=4096, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: initializer_range=0.02, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: intermediate_size=14336, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: is_llama_config=True, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: max_position_embeddings=4096, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_attention_heads=32, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_hidden_layers=32, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_key_value_heads=32, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pad_token_id=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pretraining_tp=1, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rms_norm_eps=1e-05, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_scaling=None, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_theta=10000.0, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_interleaved=False, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tie_word_embeddings=False, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: use_cache=True, +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: vocab_size=131072) +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Building model.. +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Initialize RoPE Theta = 10000.0 +12/30/2024 17:49:35 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=5|ip-26-0-163-134]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=7|ip-26-0-163-134]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=6|ip-26-0-163-134]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=4|ip-26-0-163-134]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=3|ip-26-0-163-134]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=2|ip-26-0-163-134]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=5|ip-26-0-163-134]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Total number of parameters: 8.87G (16912.25MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=7|ip-26-0-163-134]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=4|ip-26-0-163-134]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=6|ip-26-0-163-134]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=3|ip-26-0-163-134]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=2|ip-26-0-163-134]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=1|ip-26-0-163-134]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=20|ip-26-0-163-158]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=21|ip-26-0-163-158]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=23|ip-26-0-163-158]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=17|ip-26-0-163-158]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=14|ip-26-0-163-147]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=27|ip-26-0-163-220]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=22|ip-26-0-163-158]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=10|ip-26-0-163-147]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=12|ip-26-0-163-147]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=15|ip-26-0-163-147]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=13|ip-26-0-163-147]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=30|ip-26-0-163-220]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=28|ip-26-0-163-220]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=9|ip-26-0-163-147]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=26|ip-26-0-163-220]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=16|ip-26-0-163-158]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=31|ip-26-0-163-220]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=25|ip-26-0-163-220]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=19|ip-26-0-163-158]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=29|ip-26-0-163-220]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=18|ip-26-0-163-158]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=8|ip-26-0-163-147]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=20|ip-26-0-163-158]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=14|ip-26-0-163-147]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=23|ip-26-0-163-158]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=24|ip-26-0-163-220]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=21|ip-26-0-163-158]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=11|ip-26-0-163-147]: Local number of parameters: 277M (528.51MiB) +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=17|ip-26-0-163-158]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=10|ip-26-0-163-147]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=13|ip-26-0-163-147]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=15|ip-26-0-163-147]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=27|ip-26-0-163-220]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=28|ip-26-0-163-220]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=22|ip-26-0-163-158]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=30|ip-26-0-163-220]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=16|ip-26-0-163-158]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=12|ip-26-0-163-147]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=26|ip-26-0-163-220]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=19|ip-26-0-163-158]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=9|ip-26-0-163-147]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=31|ip-26-0-163-220]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=25|ip-26-0-163-220]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=29|ip-26-0-163-220]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=18|ip-26-0-163-158]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=8|ip-26-0-163-147]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=1|ip-26-0-163-134]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=24|ip-26-0-163-220]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:47 [INFO|DP=0|PP=0|TP=11|ip-26-0-163-147]: [After model building] Memory usage: 528.55MiB. Peak allocated: 5380.00MiB Peak reserved: 10550.00MiB +12/30/2024 17:49:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: No checkpoint path provided. +12/30/2024 17:49:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Parametrizing model parameters using StandardParametrizator +12/30/2024 17:49:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [Optimizer Building] Using LearningRateForSP as learning rate +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader)return f(*args, **kwargs) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs)return f(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError trainer.train(dataloader): 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( +trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs)outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) +outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + model.p2p.clear_history() + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() +trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs)raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) +Traceback (most recent call last): +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.token_position_embeddings | PP: 0/1 | Block rank: 0 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.0 | PP: 0/1 | Block rank: 1 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.1 | PP: 0/1 | Block rank: 2 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.2 | PP: 0/1 | Block rank: 3 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.3 | PP: 0/1 | Block rank: 4 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.4 | PP: 0/1 | Block rank: 5 + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.5 | PP: 0/1 | Block rank: 6 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.6 | PP: 0/1 | Block rank: 7 + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.7 | PP: 0/1 | Block rank: 8 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.8 | PP: 0/1 | Block rank: 9 + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.9 | PP: 0/1 | Block rank: 10 + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.10 | PP: 0/1 | Block rank: 11 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.11 | PP: 0/1 | Block rank: 12 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.12 | PP: 0/1 | Block rank: 13 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.13 | PP: 0/1 | Block rank: 14 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.14 | PP: 0/1 | Block rank: 15 +Traceback (most recent call last): +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.15 | PP: 0/1 | Block rank: 16 +Traceback (most recent call last): +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.16 | PP: 0/1 | Block rank: 17 + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.17 | PP: 0/1 | Block rank: 18 +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.18 | PP: 0/1 | Block rank: 19 + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.19 | PP: 0/1 | Block rank: 20 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.20 | PP: 0/1 | Block rank: 21 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.21 | PP: 0/1 | Block rank: 22 + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.22 | PP: 0/1 | Block rank: 23 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.23 | PP: 0/1 | Block rank: 24 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.24 | PP: 0/1 | Block rank: 25 + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.25 | PP: 0/1 | Block rank: 26 + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.26 | PP: 0/1 | Block rank: 27 + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.27 | PP: 0/1 | Block rank: 28 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.28 | PP: 0/1 | Block rank: 29 +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.29 | PP: 0/1 | Block rank: 30 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.30 | PP: 0/1 | Block rank: 31 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.31 | PP: 0/1 | Block rank: 32 + trainer.train(dataloader) +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.final_layer_norm | PP: 0/1 | Block rank: 33 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.lm_head | PP: 0/1 | Block rank: 34 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.cast_to_fp32 | PP: 0/1 | Block rank: 35 + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: loss | PP: 0/1 | Block rank: 36 + trainer.train(dataloader) +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Using dummy data generator + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [Training Plan] There are 1 training stages + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [Stage Stable Training Stage] start from step 1 +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: +12/30/2024 17:49:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [Start training] datetime: 2024-12-30 17:49:52.444548 | mbs: 2 | grad_accum: 8 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + return f(*args, **kwargs) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p'Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p'Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' outputs = self.pipeline_engine.train_batch_iter( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'")outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + outputs = self.pipeline_engine.train_batch_iter( +AttributeError File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( +trainer.train(dataloader) File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs)raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p'Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)model.p2p.clear_history() + +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() +raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'")model.p2p.clear_history() +Traceback (most recent call last): + +AttributeError File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + model.p2p.clear_history() + model.p2p.clear_history() + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + model.p2p.clear_history() + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( +return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + trainer.train(dataloader) +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history()outputs = self.pipeline_engine.train_batch_iter( + + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' return f(*args, **kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p'Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + return f(*args, **kwargs) + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + Traceback (most recent call last): +raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader)outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + model.p2p.clear_history() + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeErrorraise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'"): +'DistributedDataParallel' object has no attribute 'p2p'AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + + model.p2p.clear_history() + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'")outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' return f(*args, **kwargs) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + return f(*args, **kwargs) + model.p2p.clear_history() File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p'Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + return f(*args, **kwargs) + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + return f(*args, **kwargs) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + trainer.train(dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history()outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + return f(*args, **kwargs) + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + return f(*args, **kwargs) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +12/30/2024 17:49:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/30/2024 17:49:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Memory usage: 3173.06MiB. Peak allocated 5380.00MiB. Peak reserved: 12666.00MiB +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +[2024-12-30 17:50:15,438] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31317 closing signal SIGTERM +[2024-12-30 17:50:15,439] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31318 closing signal SIGTERM +[2024-12-30 17:50:15,439] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31319 closing signal SIGTERM +[2024-12-30 17:50:15,439] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31320 closing signal SIGTERM +[2024-12-30 17:50:15,439] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31322 closing signal SIGTERM +[2024-12-30 17:50:15,439] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31323 closing signal SIGTERM +[2024-12-30 17:50:15,446] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91753 closing signal SIGTERM +[2024-12-30 17:50:15,446] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91754 closing signal SIGTERM +[2024-12-30 17:50:15,446] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91755 closing signal SIGTERM +[2024-12-30 17:50:15,446] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91757 closing signal SIGTERM +[2024-12-30 17:50:15,446] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91758 closing signal SIGTERM +[2024-12-30 17:50:15,446] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91759 closing signal SIGTERM +[2024-12-30 17:50:15,446] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259844 closing signal SIGTERM +[2024-12-30 17:50:15,446] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259845 closing signal SIGTERM +[2024-12-30 17:50:15,446] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259846 closing signal SIGTERM +[2024-12-30 17:50:15,447] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259847 closing signal SIGTERM +[2024-12-30 17:50:15,446] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23706 closing signal SIGTERM +[2024-12-30 17:50:15,447] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259848 closing signal SIGTERM +[2024-12-30 17:50:15,447] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23707 closing signal SIGTERM +[2024-12-30 17:50:15,447] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259849 closing signal SIGTERM +[2024-12-30 17:50:15,447] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23708 closing signal SIGTERM +[2024-12-30 17:50:15,447] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23709 closing signal SIGTERM +[2024-12-30 17:50:15,447] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23710 closing signal SIGTERM +[2024-12-30 17:50:15,447] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23711 closing signal SIGTERM +[2024-12-30 17:50:15,447] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259850 closing signal SIGTERM +[2024-12-30 17:50:15,448] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278216 closing signal SIGTERM +[2024-12-30 17:50:15,448] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278218 closing signal SIGTERM +[2024-12-30 17:50:15,448] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278219 closing signal SIGTERM +[2024-12-30 17:50:15,449] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278220 closing signal SIGTERM +[2024-12-30 17:50:15,449] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278221 closing signal SIGTERM +[2024-12-30 17:50:15,449] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278222 closing signal SIGTERM +[2024-12-30 17:50:15,449] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278223 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114725 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114726 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114727 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114728 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114729 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114730 closing signal SIGTERM +[2024-12-30 17:50:15,451] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114731 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206033 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206035 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206036 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206037 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206038 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206039 closing signal SIGTERM +[2024-12-30 17:50:15,450] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206040 closing signal SIGTERM +[2024-12-30 17:50:15,451] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53582 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53583 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53584 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53585 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53586 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53587 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36365 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34717 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36366 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34718 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34719 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34720 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34721 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36368 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34722 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34723 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36369 closing signal SIGTERM +[2024-12-30 17:50:15,453] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36370 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190044 closing signal SIGTERM +[2024-12-30 17:50:15,453] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36371 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190046 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190047 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190048 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190049 closing signal SIGTERM +[2024-12-30 17:50:15,452] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190050 closing signal SIGTERM +[2024-12-30 17:50:15,455] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56700 closing signal SIGTERM +[2024-12-30 17:50:15,455] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56701 closing signal SIGTERM +[2024-12-30 17:50:15,455] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56702 closing signal SIGTERM +[2024-12-30 17:50:15,455] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56703 closing signal SIGTERM +[2024-12-30 17:50:15,455] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56704 closing signal SIGTERM +[2024-12-30 17:50:15,455] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56705 closing signal SIGTERM +[2024-12-30 17:50:15,455] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56706 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 538041 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 538042 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 538043 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 538044 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 538045 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915911 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915912 closing signal SIGTERM +[2024-12-30 17:50:15,457] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915913 closing signal SIGTERM +[2024-12-30 17:50:15,457] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915914 closing signal SIGTERM +[2024-12-30 17:50:15,457] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915915 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86090 closing signal SIGTERM +[2024-12-30 17:50:15,457] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915916 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86091 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86092 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86094 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86095 closing signal SIGTERM +[2024-12-30 17:50:15,456] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86096 closing signal SIGTERM +[2024-12-30 17:50:15,457] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86097 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99349 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99350 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99352 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99353 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99354 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99355 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92773 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21836 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92774 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92775 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92776 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21838 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92777 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21839 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92778 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21840 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92779 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21841 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21842 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244359 closing signal SIGTERM +[2024-12-30 17:50:15,458] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244360 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244362 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 771412 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244363 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244364 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 771413 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244365 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 771414 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 771417 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 771418 closing signal SIGTERM +[2024-12-30 17:50:15,459] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127000 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127002 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127003 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127004 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127005 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127006 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226010 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226011 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226012 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212030 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226013 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226014 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226016 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212031 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212032 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212033 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48163 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226017 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212034 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212035 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48164 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48165 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48167 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48168 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48169 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48170 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20621 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20622 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20623 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20625 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20626 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20627 closing signal SIGTERM +[2024-12-30 17:50:15,460] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 1 (pid: 107014) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58929 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260055 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58930 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58931 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260056 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260057 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260058 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58932 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260059 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260060 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58933 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58934 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58935 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130297 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490266 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130298 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130299 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490267 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490268 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490269 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130300 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490270 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130302 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490272 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130303 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492238 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492239 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492240 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492241 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492243 closing signal SIGTERM +[2024-12-30 17:50:15,461] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492244 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247661 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247662 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247663 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94129 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247664 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94130 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94131 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94132 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247665 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94133 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247666 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94134 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247668 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94135 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288555 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288557 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288558 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288559 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288560 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19376 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85900 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19377 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19379 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85901 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85902 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19380 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85903 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94603 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85904 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85906 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19381 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85907 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94604 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94605 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94606 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94607 closing signal SIGTERM +[2024-12-30 17:50:15,462] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19382 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94608 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255549 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255550 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255551 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255552 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255553 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255554 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892866 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892867 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892868 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892869 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892870 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892872 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39005 closing signal SIGTERM +[2024-12-30 17:50:15,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56380 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39006 closing signal SIGTERM +[2024-12-30 17:50:15,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56381 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39008 closing signal SIGTERM +[2024-12-30 17:50:15,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56382 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163118 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39010 closing signal SIGTERM +[2024-12-30 17:50:15,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56383 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163119 closing signal SIGTERM +[2024-12-30 17:50:15,463] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39011 closing signal SIGTERM +[2024-12-30 17:50:15,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56384 closing signal SIGTERM +[2024-12-30 17:50:15,464] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163120 closing signal SIGTERM +[2024-12-30 17:50:15,465] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163121 closing signal SIGTERM +[2024-12-30 17:50:15,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56386 closing signal SIGTERM +[2024-12-30 17:50:15,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56387 closing signal SIGTERM +[2024-12-30 17:50:15,465] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163122 closing signal SIGTERM +[2024-12-30 17:50:15,465] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163124 closing signal SIGTERM +[2024-12-30 17:50:15,465] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944142 closing signal SIGTERM +[2024-12-30 17:50:15,465] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944143 closing signal SIGTERM +[2024-12-30 17:50:15,465] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944144 closing signal SIGTERM +[2024-12-30 17:50:15,465] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944145 closing signal SIGTERM +[2024-12-30 17:50:15,465] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944146 closing signal SIGTERM +[2024-12-30 17:50:15,466] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944148 closing signal SIGTERM +[2024-12-30 17:50:15,467] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 69313) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2024-12-30 17:50:15,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162231 closing signal SIGTERM +[2024-12-30 17:50:15,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162232 closing signal SIGTERM +[2024-12-30 17:50:15,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162233 closing signal SIGTERM +[2024-12-30 17:50:15,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162234 closing signal SIGTERM +[2024-12-30 17:50:15,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162235 closing signal SIGTERM +[2024-12-30 17:50:15,468] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162237 closing signal SIGTERM +[2024-12-30 17:50:15,471] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 207995 closing signal SIGTERM +[2024-12-30 17:50:15,471] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 207996 closing signal SIGTERM +[2024-12-30 17:50:15,471] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 207997 closing signal SIGTERM +[2024-12-30 17:50:15,471] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 207998 closing signal SIGTERM +[2024-12-30 17:50:15,471] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 207999 closing signal SIGTERM +[2024-12-30 17:50:15,471] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 208001 closing signal SIGTERM +[2024-12-30 17:50:15,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56853 closing signal SIGTERM +[2024-12-30 17:50:15,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56854 closing signal SIGTERM +[2024-12-30 17:50:15,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56855 closing signal SIGTERM +[2024-12-30 17:50:15,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56856 closing signal SIGTERM +[2024-12-30 17:50:15,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56858 closing signal SIGTERM +[2024-12-30 17:50:15,470] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56859 closing signal SIGTERM +[2024-12-30 17:50:15,473] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79045 closing signal SIGTERM +[2024-12-30 17:50:15,473] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79046 closing signal SIGTERM +[2024-12-30 17:50:15,473] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79047 closing signal SIGTERM +[2024-12-30 17:50:15,473] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79048 closing signal SIGTERM +[2024-12-30 17:50:15,473] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79049 closing signal SIGTERM +[2024-12-30 17:50:15,473] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79050 closing signal SIGTERM +[2024-12-30 17:50:15,473] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79051 closing signal SIGTERM +[2024-12-30 17:50:15,519] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_1yn_hbjr/14016402_9b_7eoif/attempt_0/1/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[0]: + time : 2024-12-30_17:49:54 + host : ip-26-0-163-134.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 107013) + error_file: /tmp/torchelastic_1yn_hbjr/14016402_9b_7eoif/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[2]: + time : 2024-12-30_17:49:52 + host : ip-26-0-163-134.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 107015) + error_file: /tmp/torchelastic_1yn_hbjr/14016402_9b_7eoif/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[3]: + time : 2024-12-30_17:49:52 + host : ip-26-0-163-134.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 107016) + error_file: /tmp/torchelastic_1yn_hbjr/14016402_9b_7eoif/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[4]: + time : 2024-12-30_17:49:52 + host : ip-26-0-163-134.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 107017) + error_file: /tmp/torchelastic_1yn_hbjr/14016402_9b_7eoif/attempt_0/4/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[5]: + time : 2024-12-30_17:49:52 + host : ip-26-0-163-134.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 107018) + error_file: /tmp/torchelastic_1yn_hbjr/14016402_9b_7eoif/attempt_0/5/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[6]: + time : 2024-12-30_17:49:52 + host : ip-26-0-163-134.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 107019) + error_file: /tmp/torchelastic_1yn_hbjr/14016402_9b_7eoif/attempt_0/6/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[7]: + time : 2024-12-30_17:49:52 + host : ip-26-0-163-134.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 107020) + error_file: /tmp/torchelastic_1yn_hbjr/14016402_9b_7eoif/attempt_0/7/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +------------------------------------------------------------ +Root Cause (first observed failure): +[1]: + time : 2024-12-30_17:49:52 + host : ip-26-0-163-134.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 107014) + error_file: /tmp/torchelastic_1yn_hbjr/14016402_9b_7eoif/attempt_0/1/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +============================================================ +[2024-12-30 17:50:15,530] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_3ac4z5kw/14016402_gv_gdmu1/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2024-12-30_17:49:52 + host : ip-26-0-175-34.ec2.internal + rank : 505 (local_rank: 1) + exitcode : 1 (pid: 69314) + error_file: /tmp/torchelastic_3ac4z5kw/14016402_gv_gdmu1/attempt_0/1/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[2]: + time : 2024-12-30_17:49:52 + host : ip-26-0-175-34.ec2.internal + rank : 506 (local_rank: 2) + exitcode : 1 (pid: 69315) + error_file: /tmp/torchelastic_3ac4z5kw/14016402_gv_gdmu1/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[3]: + time : 2024-12-30_17:49:52 + host : ip-26-0-175-34.ec2.internal + rank : 507 (local_rank: 3) + exitcode : 1 (pid: 69316) + error_file: /tmp/torchelastic_3ac4z5kw/14016402_gv_gdmu1/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[4]: + time : 2024-12-30_17:49:52 + host : ip-26-0-175-34.ec2.internal + rank : 508 (local_rank: 4) + exitcode : 1 (pid: 69317) + error_file: /tmp/torchelastic_3ac4z5kw/14016402_gv_gdmu1/attempt_0/4/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[5]: + time : 2024-12-30_17:49:52 + host : ip-26-0-175-34.ec2.internal + rank : 509 (local_rank: 5) + exitcode : 1 (pid: 69318) + error_file: /tmp/torchelastic_3ac4z5kw/14016402_gv_gdmu1/attempt_0/5/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[6]: + time : 2024-12-30_17:49:52 + host : ip-26-0-175-34.ec2.internal + rank : 510 (local_rank: 6) + exitcode : 1 (pid: 69319) + error_file: /tmp/torchelastic_3ac4z5kw/14016402_gv_gdmu1/attempt_0/6/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[7]: + time : 2024-12-30_17:49:52 + host : ip-26-0-175-34.ec2.internal + rank : 511 (local_rank: 7) + exitcode : 1 (pid: 69320) + error_file: /tmp/torchelastic_3ac4z5kw/14016402_gv_gdmu1/attempt_0/7/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-30_17:49:52 + host : ip-26-0-175-34.ec2.internal + rank : 504 (local_rank: 0) + exitcode : 1 (pid: 69313) + error_file: /tmp/torchelastic_3ac4z5kw/14016402_gv_gdmu1/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +============================================================ +srun: error: ip-26-0-163-134: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=14016402.0 +[2024-12-30 17:50:15,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58930 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58931 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244359 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58932 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244360 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58933 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244362 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58934 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244363 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58935 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244364 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 244365 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56853 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 771412 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 771413 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56854 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 771414 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 771417 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56855 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56856 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 771418 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 538041 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 538042 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226010 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 538043 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56858 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 538044 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56859 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226011 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247662 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 538045 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226012 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226013 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247663 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226014 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247664 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247665 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226016 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23706 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 226017 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23707 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247666 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23708 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 247668 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23709 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23710 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915911 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23711 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915912 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421677 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91753 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915913 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421678 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91754 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421679 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91755 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53607 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915914 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39005 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421680 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56380 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91757 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915915 closing signal SIGTERM +[2024-12-30 17:50:15,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39006 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39008 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421681 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56381 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91758 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2915916 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39010 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421682 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91759 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53608 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 39011 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56382 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56383 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53609 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421683 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94129 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56384 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421684 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278216 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94130 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94131 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53610 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53582 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278218 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56386 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85901 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94132 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56387 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278219 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53611 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53583 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85902 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85903 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94133 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94134 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53584 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85904 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278220 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53612 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85906 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189886 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53585 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94135 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490266 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278221 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53614 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 85907 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189887 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490267 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490268 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53586 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92773 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189888 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278222 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490269 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 278223 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490270 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56700 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 53587 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92774 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189890 closing signal SIGTERM +[2024-12-30 17:50:15,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163118 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56701 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 490272 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92776 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288555 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189891 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34717 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163119 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 150912 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56702 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 189892 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56703 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892866 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288557 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892867 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92777 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288558 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163120 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 150914 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288559 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56704 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892868 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288560 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34718 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56705 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892869 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 56706 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212030 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92778 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34719 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163121 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 150915 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212031 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892870 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 92779 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212032 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 892872 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36365 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212033 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48163 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34720 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163122 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 150916 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 150917 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48164 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163124 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 150918 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212034 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 298019 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36366 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48165 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 150919 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212035 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36368 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34721 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22404 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 298020 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36369 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190044 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34722 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 207995 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36370 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48167 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22405 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 207996 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190046 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48168 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26867 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 207997 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 298021 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190047 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34723 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 207998 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 298022 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 36371 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190048 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22406 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 207999 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206033 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190049 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48169 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22407 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 208001 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 48170 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 298023 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206035 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 298024 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 190050 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19376 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26868 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22408 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206036 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22409 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 326752 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22410 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19377 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 298025 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26869 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 298026 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130297 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206037 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 326753 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206038 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19379 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255549 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130298 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26871 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 326755 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711155 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206039 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255550 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711156 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130299 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19380 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255551 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259844 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711157 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255552 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 326756 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711158 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26872 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206040 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169763 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711159 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130300 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19381 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255553 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259845 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711160 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 647354 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 255554 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 326758 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169764 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26873 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20621 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260055 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259846 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 647355 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130302 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 19382 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169765 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711161 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20622 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260056 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 711162 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 647357 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20623 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259847 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260057 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 130303 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20625 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169766 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20626 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169767 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 647359 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259848 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169768 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 647360 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260058 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259849 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 647361 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 20627 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260059 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 259850 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169769 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94603 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 169770 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492238 closing signal SIGTERM +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3090903 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94604 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 260060 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3090904 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94605 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94606 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492239 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3090905 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94607 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492240 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 94608 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31317 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3090907 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492241 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114555 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99349 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31318 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3090908 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114556 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492243 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3090909 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114725 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114557 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2492244 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79046 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99350 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114558 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79047 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31319 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3090910 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114726 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114727 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162231 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127000 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114728 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114559 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79048 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34112 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114560 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21836 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99353 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162232 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79049 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31320 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114561 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162233 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127002 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31322 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114729 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34113 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162234 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127003 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31323 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114730 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34114 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21838 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55451 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99354 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162235 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127004 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79050 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34115 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944142 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127005 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79051 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114731 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114562 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21839 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944143 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 99355 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162237 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 127006 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86091 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34116 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55452 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944144 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34117 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55453 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2664252 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944145 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86092 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34118 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21840 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86094 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86095 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55454 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944146 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86096 closing signal SIGTERM +[2024-12-30 17:50:15,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1944148 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21841 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2664253 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 21842 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55455 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 86097 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2664254 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2664255 closing signal SIGTERM +[2024-12-30 17:50:15,880] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 55457 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2664256 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2664257 closing signal SIGTERM +[2024-12-30 17:50:15,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2664258 closing signal SIGTERM +srun: error: ip-26-0-175-34: task 61: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 288480 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 537965 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 244280 got signal: 15 +srun: error: ip-26-0-169-86: task 27: Exited with exit code 1 +srun: error: ip-26-0-163-226: task 4: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 36289 got signal: 15 +srun: error: ip-26-0-163-220: task 3: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + sys.exit(main()) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + self.close() # terminate all running procs + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 99272 got signal: 15 + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 38929 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 19298 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 23631 got signal: 15 +srun: error: ip-26-0-169-132: task 28: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = agent.run() + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 20544 got signal: 15 + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1944066 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 126925 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 94526 got signal: 15 +srun: error: ip-26-0-170-160: task 36: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 31240 got signal: 15 +srun: error: ip-26-0-171-230: task 43: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 771337 got signal: 15 +srun: error: ip-26-0-164-187: task 10: Exited with exit code 1 +srun: error: ip-26-0-172-116: task 47: Exited with exit code 1 +srun: error: ip-26-0-164-75: task 9: Exited with exit code 1 +srun: error: ip-26-0-175-132: task 62: Exited with exit code 1 +srun: error: ip-26-0-174-36: task 55: Exited with exit code 1 +srun: error: ip-26-0-174-100: task 56: Exited with exit code 1 +srun: error: ip-26-0-173-202: task 53: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 211954 got signal: 15 +srun: error: ip-26-0-164-45: task 8: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 163041 got signal: 15 + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 114649 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 91677 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 255475 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 259979 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper +srun: error: ip-26-0-171-62: task 39: Exited with exit code 1 + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 207920 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 490190 got signal: 15 +srun: error: ip-26-0-168-34: task 22: Exited with exit code 1 +srun: error: ip-26-0-166-15: task 20: Exited with exit code 1 +srun: error: ip-26-0-172-252: task 50: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + sys.exit(main()) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + run_result = self._monitor_workers(self._worker_group) + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 86016 got signal: 15 + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 205856 got signal: 15 + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2492162 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 130220 got signal: 15 +srun: error: ip-26-0-171-249: task 44: Exited with exit code 1 +srun: error: ip-26-0-174-186: task 57: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + elastic_launch( + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 892789 got signal: 15 + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 53505 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +srun: error: ip-26-0-169-247: task 32: Exited with exit code 1 + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 150837 got signal: 15 + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 162156 got signal: 15 +srun: error: ip-26-0-175-19: task 60: Exited with exit code 1 +srun: error: ip-26-0-171-102: task 41: Exited with exit code 1 +srun: error: ip-26-0-173-121: task 52: Exited with exit code 1 +srun: error: ip-26-0-165-202: task 18: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 58854 got signal: 15 +srun: error: ip-26-0-165-59: task 15: Exited with exit code 1 +srun: error: ip-26-0-165-38: task 14: Exited with exit code 1 +srun: error: ip-26-0-170-31: task 33: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 22329 got signal: 15 +srun: error: ip-26-0-171-88: task 40: Exited with exit code 1 +srun: error: ip-26-0-172-147: task 49: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 278141 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 94054 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 189969 got signal: 15 +srun: error: ip-26-0-163-147: task 1: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 56624 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 247586 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 78971 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 326675 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2915835 got signal: 15 +srun: error: ip-26-0-169-207: task 30: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 56777 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 56303 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 189809 got signal: 15 +srun: error: ip-26-0-168-30: task 21: Exited with exit code 1 +srun: error: ip-26-0-165-213: task 19: Exited with exit code 1 +srun: error: ip-26-0-164-0: task 6: Exited with exit code 1 +srun: error: ip-26-0-171-21: task 37: Exited with exit code 1 +srun: error: ip-26-0-169-139: task 29: Exited with exit code 1 +srun: error: ip-26-0-174-196: task 58: Exited with exit code 1 +srun: error: ip-26-0-170-143: task 35: Exited with exit code 1 +srun: error: ip-26-0-164-18: task 7: Exited with exit code 1 +[2024-12-30 17:50:20,036] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-56.ec2.internal_34641_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +[2024-12-30 17:50:20,045] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-57.ec2.internal_169688_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +[2024-12-30 17:50:20,059] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-168.ec2.internal_114481_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run +[2024-12-30 17:50:20,061] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-24.ec2.internal_225934_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 225934 got signal: 15 + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 48087 got signal: 15 +srun: error: ip-26-0-165-131: task 16: Exited with exit code 1 +[2024-12-30 17:50:20,092] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-173-7.ec2.internal_34037_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +[2024-12-30 17:50:20,096] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-163-158.ec2.internal_53532_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +[2024-12-30 17:50:20,174] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-207.ec2.internal_85825_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-165-164: task 17: Exited with exit code 1 +[2024-12-30 17:50:20,201] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-173-246.ec2.internal_3090826_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler +srun: error: ip-26-0-163-236: task 5: Exited with exit code 1 + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 92696 got signal: 15 +[2024-12-30 17:50:20,248] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-73.ec2.internal_259769_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +[2024-12-30 17:50:20,252] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-120.ec2.internal_297944_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +[2024-12-30 17:50:20,255] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-142.ec2.internal_21761_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +[2024-12-30 17:50:20,272] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-170-132.ec2.internal_647278_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +[2024-12-30 17:50:20,274] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-175-165.ec2.internal_2664178_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +[2024-12-30 17:50:20,279] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-238.ec2.internal_26790_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2664178 got signal: 15 +[2024-12-30 17:50:20,292] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-236.ec2.internal_421601_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +[2024-12-30 17:50:20,293] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-174-240.ec2.internal_55374_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 34037 got signal: 15 +srun: error: ip-26-0-165-24: task 13: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 85825 got signal: 15 +[2024-12-30 17:50:20,383] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-239.ec2.internal_711080_0' has failed to send a keep-alive heartbeat to the rendezvous '14016402' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 297944 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 421601 got signal: 15 + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 55374 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 53532 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 21761 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 26790 got signal: 15 +srun: error: ip-26-0-168-52: task 23: Exited with exit code 1 +srun: error: ip-26-0-175-165: task 63: Exited with exit code 1 +srun: error: ip-26-0-168-95: task 24: Exited with exit code 1 +srun: error: ip-26-0-173-7: task 51: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 114481 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main +srun: error: ip-26-0-174-240: task 59: Exited with exit code 1 + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 169688 got signal: 15 +srun: error: ip-26-0-172-142: task 48: Exited with exit code 1 +srun: error: ip-26-0-168-120: task 25: Exited with exit code 1 +srun: error: ip-26-0-164-236: task 12: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main +srun: error: ip-26-0-163-158: task 2: Exited with exit code 1 + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 647278 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 3090826 got signal: 15 +srun: error: ip-26-0-164-207: task 11: Exited with exit code 1 +srun: error: ip-26-0-168-238: task 26: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 711080 got signal: 15 +srun: error: ip-26-0-171-168: task 42: Exited with exit code 1 +srun: error: ip-26-0-172-57: task 45: Exited with exit code 1 +srun: error: ip-26-0-170-132: task 34: Exited with exit code 1 +srun: error: ip-26-0-173-246: task 54: Exited with exit code 1 +srun: error: ip-26-0-169-239: task 31: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 259769 got signal: 15 +srun: error: ip-26-0-172-73: task 46: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 34641 got signal: 15 +srun: error: ip-26-0-171-56: task 38: Exited with exit code 1 diff --git a/logs/14016454-bench_1.34G_dp2_tp32_pp8_acc16_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/14016454-bench_1.34G_dp2_tp32_pp8_acc16_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..77bb14bf9c448f8b36e7a8445317314f980aedd6 --- /dev/null +++ b/logs/14016454-bench_1.34G_dp2_tp32_pp8_acc16_mbs8_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,5481 @@ ++ '[' -z 14016454 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/a846435528b4b760494a836f96f0739889253530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-163-[220,226,236],ip-26-0-164-[0,18,45,187,207,236],ip-26-0-165-[24,38,59,131,164,202,213],ip-26-0-166-[15,36,125,214,244],ip-26-0-167-[9,51,175,177,217],ip-26-0-168-[30,34,52,95,120,238],ip-26-0-169-[86,139,207],ip-26-0-170-[143,160],ip-26-0-171-[21,56,62,88,102,168,230,249],ip-26-0-172-[57,73,116,142,147,252],ip-26-0-173-[7,121,202,246],ip-26-0-174-[36,100,186],ip-26-0-175-[19,34,132,165,170,241]' ++ export 'NODELIST=ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-167-217 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165 +ip-26-0-175-170 +ip-26-0-175-241' ++ NODELIST='ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-167-217 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165 +ip-26-0-175-170 +ip-26-0-175-241' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-163-[220,226,236],ip-26-0-164-[0,18,45,187,207,236],ip-26-0-165-[24,38,59,131,164,202,213],ip-26-0-166-[15,36,125,214,244],ip-26-0-167-[9,51,175,177,217],ip-26-0-168-[30,34,52,95,120,238],ip-26-0-169-[86,139,207],ip-26-0-170-[143,160],ip-26-0-171-[21,56,62,88,102,168,230,249],ip-26-0-172-[57,73,116,142,147,252],ip-26-0-173-[7,121,202,246],ip-26-0-174-[36,100,186],ip-26-0-175-[19,34,132,165,170,241]' ++ export MASTER_NODE=ip-26-0-163-220 ++ MASTER_NODE=ip-26-0-163-220 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=64 ++ NNODES=64 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=512 ++ WORLD_SIZE=512 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-43 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 52-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 52-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 52-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 52-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-163-220' +Master node: ip-26-0-163-220 ++ echo 'All nodes: ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-167-217 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165 +ip-26-0-175-170 +ip-26-0-175-241' +All nodes: ip-26-0-163-220 +ip-26-0-163-226 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-167-217 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165 +ip-26-0-175-170 +ip-26-0-175-241 ++ echo 'World size: 512' +World size: 512 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=64 --nproc_per_node=8 --rdzv_id=14016454 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-163-220:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.34G_dp2_tp32_pp8_acc16_mbs8_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2024-12-31 00:37:59,224] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,224] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,224] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,224] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,224] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,223] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,223] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,223] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,223] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,223] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,224] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,224] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,225] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,227] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,226] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,228] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,228] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,228] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,228] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,228] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,228] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,229] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,229] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,230] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,231] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,234] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,235] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,235] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,235] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,235] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,234] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,234] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,234] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,234] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,234] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,234] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,234] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,234] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,234] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,234] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,235] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,235] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,235] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,235] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,235] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,236] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,236] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,236] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,236] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,236] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,236] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,236] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,236] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,236] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,236] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,238] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,238] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,238] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,238] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,238] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,239] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,239] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,239] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,239] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,239] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,239] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,239] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,239] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,239] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,239] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,240] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,241] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,241] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,241] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,241] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,244] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,244] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,244] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,244] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,244] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,248] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,248] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,248] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,248] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,248] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,249] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,249] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,249] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,249] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,249] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,252] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,252] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,252] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,252] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,265] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,265] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,265] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,265] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,265] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,228] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,228] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,228] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,228] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,271] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,271] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,271] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,271] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,271] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,275] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,276] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,276] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,276] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,276] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,282] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,282] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,282] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,282] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,282] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,294] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,294] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,294] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,294] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,294] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,299] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,299] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,299] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,299] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,299] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,308] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,310] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,310] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,310] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,310] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,310] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,316] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,316] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,316] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,316] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,316] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,308] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,308] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,308] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,308] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,360] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,360] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,360] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,360] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,360] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,360] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,360] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,360] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,360] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,360] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,382] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,382] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,382] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,382] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,382] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,391] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,391] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,391] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,391] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,391] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,409] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,409] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,409] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,409] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,409] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,464] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:37:59,464] torch.distributed.run: [WARNING] +[2024-12-31 00:37:59,464] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:37:59,464] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:37:59,464] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,766] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:38:02,769] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:38:02,766] torch.distributed.run: [WARNING] +[2024-12-31 00:38:02,766] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,766] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:38:02,766] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,769] torch.distributed.run: [WARNING] +[2024-12-31 00:38:02,769] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,769] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:38:02,769] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,770] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:38:02,766] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:38:02,767] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,770] torch.distributed.run: [WARNING] +[2024-12-31 00:38:02,770] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,770] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:38:02,770] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,766] torch.distributed.run: [WARNING] +[2024-12-31 00:38:02,766] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,766] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:38:02,766] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,922] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-31 00:38:02,922] torch.distributed.run: [WARNING] +[2024-12-31 00:38:02,922] torch.distributed.run: [WARNING] ***************************************** +[2024-12-31 00:38:02,922] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-31 00:38:02,922] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Config: +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Config(general=GeneralArgs(project='debug', +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: run='1.34G_dp2_tp32_pp8_acc16_mbs8_seq4096_zero1_tpmodeRED_vocab131k', +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: seed=42, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: step=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: consumed_train_samples=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: ignore_sanity_checks=True), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: parallelism=ParallelismArgs(dp=2, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: pp=8, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: tp=32, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: pp_engine=, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: tp_mode=, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: tp_linear_async_communication=True, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: recompute_layer=False, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: tp_recompute_allgather=True, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: expert_parallel_size=1), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: eos_token_id=0, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: hidden_act='silu', +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: hidden_size=2048, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: initializer_range=0.02, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: intermediate_size=8192, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: is_llama_config=True, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: max_position_embeddings=4096, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: num_attention_heads=32, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: num_hidden_layers=16, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: num_key_value_heads=32, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: pad_token_id=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: pretraining_tp=1, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: rms_norm_eps=1e-05, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: rope_scaling=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: rope_theta=10000.0, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: rope_interleaved=False, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: tie_word_embeddings=True, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: use_cache=True, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: vocab_size=131072), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: init_method=RandomInit(std=0.02), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: dtype=torch.bfloat16, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: make_vocab_size_divisible_by=1, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: ddp_bucket_cap_mb=25), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: tokenizer_revision=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: tokenizer_max_length=None), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: checkpoint_interval=10000, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: save_initial_state=False, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: save_final_state=False, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: resume_checkpoint_path=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: checkpoints_path_is_shared_file_system=False), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: logging=LoggingArgs(log_level='info', +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: log_level_replica='info', +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: iteration_step_info_interval=1), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: tokens=TokensArgs(sequence_length=4096, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: train_steps=100, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: micro_batch_size=8, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: batch_accumulation_per_replica=16, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: val_check_interval=100, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: limit_val_batches=0, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: limit_test_batches=0), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: adam_beta1=0.9, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: adam_beta2=0.95, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: torch_adam_is_fused=True, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: name='adamW'), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: zero_stage=1, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: weight_decay=0.01, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: clip_grad=1.0, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: accumulate_grad_in_fp32=True, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: lr_warmup_steps=2, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: lr_warmup_style='linear', +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: lr_decay_style='cosine', +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: lr_decay_steps=13, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: lr_decay_starting_step=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: min_decay_lr=1e-05)), +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: start_training_step=1, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: data=DataArgs(dataset=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: seed=42, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: num_loading_workers=1))], +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: profiler=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: lighteval=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: s3_upload=None) +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Model Config: +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: LlamaConfig(bos_token_id=0, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: eos_token_id=0, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: hidden_act='silu', +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: hidden_size=2048, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: initializer_range=0.02, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: intermediate_size=8192, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: is_llama_config=True, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: max_position_embeddings=4096, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: num_attention_heads=32, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: num_hidden_layers=16, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: num_key_value_heads=32, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: pad_token_id=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: pretraining_tp=1, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: rms_norm_eps=1e-05, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: rope_scaling=None, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: rope_theta=10000.0, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: rope_interleaved=False, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: tie_word_embeddings=True, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: use_cache=True, +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: vocab_size=131072) +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Building model.. +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Initialize RoPE Theta = 10000.0 +12/31/2024 00:39:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=5|ip-26-0-163-220]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=5|ip-26-0-169-139]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=5|ip-26-0-166-125]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=5|ip-26-0-172-252]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=5|ip-26-0-171-230]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=5|ip-26-0-164-45]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=5|ip-26-0-163-220]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=5|ip-26-0-169-139]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=5|ip-26-0-166-125]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=5|ip-26-0-172-252]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=5|ip-26-0-171-230]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=5|ip-26-0-164-45]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=7|ip-26-0-166-125]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=7|ip-26-0-169-139]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=7|ip-26-0-172-252]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=7|ip-26-0-163-220]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=7|ip-26-0-166-125]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=7|ip-26-0-169-139]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=7|ip-26-0-172-252]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=7|ip-26-0-174-186]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=7|ip-26-0-163-220]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=7|ip-26-0-174-186]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Total number of parameters: 1.61G (3076.12MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=0|ip-26-0-169-139]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=5|ip-26-0-174-186]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=0|ip-26-0-166-125]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=1|ip-26-0-163-220]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=0|ip-26-0-164-45]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=5|ip-26-0-167-51]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=0|ip-26-0-172-252]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=0|ip-26-0-169-139]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=5|ip-26-0-174-186]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: No checkpoint path provided. +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=0|ip-26-0-164-45]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=0|ip-26-0-166-125]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=1|ip-26-0-163-220]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Parametrizing model parameters using StandardParametrizator +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=0|ip-26-0-172-252]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=5|ip-26-0-167-51]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=7|ip-26-0-171-230]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=7|ip-26-0-167-51]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=6|ip-26-0-163-220]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=7|ip-26-0-171-230]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=7|ip-26-0-167-51]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=6|ip-26-0-163-220]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=7|ip-26-0-164-45]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=15|ip-26-0-163-226]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=7|ip-26-0-164-45]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=15|ip-26-0-163-226]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=2|ip-26-0-163-220]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=15|ip-26-0-169-207]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=3|ip-26-0-163-220]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=3|ip-26-0-166-125]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=3|ip-26-0-169-139]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=1|ip-26-0-172-252]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=0|ip-26-0-171-230]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=3|ip-26-0-172-252]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=2|ip-26-0-163-220]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=1|ip-26-0-169-139]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=15|ip-26-0-169-207]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=3|ip-26-0-163-220]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=1|ip-26-0-166-125]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=1|ip-26-0-164-45]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=3|ip-26-0-166-125]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=1|ip-26-0-172-252]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=1|ip-26-0-171-230]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=0|ip-26-0-167-51]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=0|ip-26-0-174-186]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=3|ip-26-0-169-139]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=3|ip-26-0-174-186]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=0|ip-26-0-171-230]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=3|ip-26-0-172-252]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=1|ip-26-0-169-139]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=1|ip-26-0-166-125]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=1|ip-26-0-164-45]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=30|ip-26-0-164-0]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=1|ip-26-0-171-230]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=0|ip-26-0-167-51]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=6|ip-26-0-166-125]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=6|ip-26-0-167-51]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=0|ip-26-0-174-186]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=6|ip-26-0-164-45]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=27|ip-26-0-164-0]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=6|ip-26-0-169-139]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=30|ip-26-0-164-0]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=6|ip-26-0-172-252]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=6|ip-26-0-174-186]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=6|ip-26-0-171-230]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=3|ip-26-0-174-186]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=6|ip-26-0-166-125]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=19|ip-26-0-163-236]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=6|ip-26-0-167-51]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=27|ip-26-0-164-0]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=6|ip-26-0-164-45]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=6|ip-26-0-169-139]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=6|ip-26-0-172-252]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=6|ip-26-0-174-186]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=6|ip-26-0-171-230]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=19|ip-26-0-169-86]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=12|ip-26-0-163-226]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=19|ip-26-0-163-236]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 9500.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=25|ip-26-0-164-0]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=16|ip-26-0-163-236]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=15|ip-26-0-172-57]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=15|ip-26-0-166-15]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=19|ip-26-0-169-86]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=12|ip-26-0-163-226]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=16|ip-26-0-163-236]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 9500.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=15|ip-26-0-172-57]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=15|ip-26-0-166-15]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=8|ip-26-0-163-226]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=25|ip-26-0-164-0]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=19|ip-26-0-166-214]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=19|ip-26-0-172-73]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=2|ip-26-0-172-252]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=2|ip-26-0-174-186]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=2|ip-26-0-169-139]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=2|ip-26-0-166-125]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=17|ip-26-0-163-236]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=2|ip-26-0-171-230]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=8|ip-26-0-163-226]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=25|ip-26-0-170-143]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=19|ip-26-0-166-214]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=16|ip-26-0-169-86]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=19|ip-26-0-172-73]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=2|ip-26-0-172-252]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=2|ip-26-0-174-186]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=3|ip-26-0-164-45]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=2|ip-26-0-169-139]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=2|ip-26-0-166-125]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=3|ip-26-0-171-230]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=3|ip-26-0-167-51]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=17|ip-26-0-163-236]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 9500.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=2|ip-26-0-171-230]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=25|ip-26-0-170-143]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=1|ip-26-0-167-51]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=19|ip-26-0-171-56]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=24|ip-26-0-170-143]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=19|ip-26-0-175-132]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=16|ip-26-0-169-86]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=19|ip-26-0-168-120]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=24|ip-26-0-164-0]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=3|ip-26-0-171-230]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=16|ip-26-0-172-73]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=3|ip-26-0-164-45]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=3|ip-26-0-167-51]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=16|ip-26-0-166-214]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=25|ip-26-0-166-244]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=4|ip-26-0-163-220]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=4|ip-26-0-169-139]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=1|ip-26-0-167-51]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=4|ip-26-0-166-125]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=19|ip-26-0-171-56]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=19|ip-26-0-175-132]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 9494.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=19|ip-26-0-168-120]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=4|ip-26-0-172-252]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=24|ip-26-0-170-143]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=24|ip-26-0-164-0]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=16|ip-26-0-172-73]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=4|ip-26-0-167-51]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=8|ip-26-0-169-207]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=23|ip-26-0-163-236]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=25|ip-26-0-166-244]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=4|ip-26-0-163-220]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=1|ip-26-0-174-186]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=16|ip-26-0-166-214]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=4|ip-26-0-169-139]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=20|ip-26-0-163-236]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=4|ip-26-0-166-125]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=4|ip-26-0-172-252]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=13|ip-26-0-163-226]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=19|ip-26-0-165-164]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=8|ip-26-0-172-57]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=9|ip-26-0-163-226]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=4|ip-26-0-174-186]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=4|ip-26-0-167-51]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=8|ip-26-0-166-15]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=16|ip-26-0-165-164]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=23|ip-26-0-163-236]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 9500.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=16|ip-26-0-171-56]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=16|ip-26-0-175-132]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=8|ip-26-0-172-57]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=9|ip-26-0-163-226]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=8|ip-26-0-169-207]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=13|ip-26-0-163-226]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=16|ip-26-0-168-120]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=25|ip-26-0-168-238]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=20|ip-26-0-163-236]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 9500.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=8|ip-26-0-166-15]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=25|ip-26-0-171-62]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=1|ip-26-0-174-186]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=4|ip-26-0-174-186]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=25|ip-26-0-165-202]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=16|ip-26-0-171-56]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=25|ip-26-0-173-121]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=16|ip-26-0-175-132]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 9494.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=19|ip-26-0-165-164]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=25|ip-26-0-168-238]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=16|ip-26-0-168-120]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=16|ip-26-0-165-164]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=22|ip-26-0-163-236]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=25|ip-26-0-171-62]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=15|ip-26-0-171-249]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=25|ip-26-0-165-202]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=22|ip-26-0-163-236]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 9500.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=25|ip-26-0-173-121]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=15|ip-26-0-171-249]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=15|ip-26-0-165-131]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=18|ip-26-0-163-236]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=2|ip-26-0-167-51]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=2|ip-26-0-164-45]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=15|ip-26-0-165-131]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=18|ip-26-0-163-236]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 9500.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=2|ip-26-0-164-45]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=2|ip-26-0-167-51]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=30|ip-26-0-170-143]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=27|ip-26-0-170-143]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=10|ip-26-0-163-226]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=26|ip-26-0-164-0]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=11|ip-26-0-163-226]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=29|ip-26-0-164-0]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=30|ip-26-0-170-143]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=27|ip-26-0-170-143]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=4|ip-26-0-164-45]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=10|ip-26-0-163-226]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=28|ip-26-0-164-0]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=26|ip-26-0-164-0]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=29|ip-26-0-164-0]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=4|ip-26-0-164-45]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=12|ip-26-0-169-207]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=28|ip-26-0-164-0]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=11|ip-26-0-163-226]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=15|ip-26-0-167-9]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=12|ip-26-0-169-207]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=15|ip-26-0-167-9]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=27|ip-26-0-166-244]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=4|ip-26-0-171-230]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=15|ip-26-0-174-36]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=8|ip-26-0-167-9]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=24|ip-26-0-173-121]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=12|ip-26-0-166-15]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=30|ip-26-0-166-244]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=9|ip-26-0-169-207]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=17|ip-26-0-169-86]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=12|ip-26-0-167-9]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=17|ip-26-0-165-164]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=31|ip-26-0-164-0]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=4|ip-26-0-171-230]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=27|ip-26-0-165-202]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=9|ip-26-0-172-57]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=8|ip-26-0-167-9]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=17|ip-26-0-166-214]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=15|ip-26-0-174-36]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=24|ip-26-0-173-121]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=9|ip-26-0-169-207]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=12|ip-26-0-166-15]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=27|ip-26-0-166-244]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=30|ip-26-0-166-244]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=31|ip-26-0-164-0]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=12|ip-26-0-167-9]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=17|ip-26-0-169-86]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=17|ip-26-0-165-164]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=27|ip-26-0-165-202]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=24|ip-26-0-166-244]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=24|ip-26-0-168-238]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=17|ip-26-0-166-214]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=24|ip-26-0-171-62]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=30|ip-26-0-165-202]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=9|ip-26-0-172-57]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=17|ip-26-0-171-56]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=17|ip-26-0-172-73]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=9|ip-26-0-166-15]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=9|ip-26-0-171-249]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=27|ip-26-0-173-121]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=21|ip-26-0-163-236]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=24|ip-26-0-166-244]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=24|ip-26-0-168-238]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=14|ip-26-0-163-226]: Local number of parameters: 14.7M (28.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=30|ip-26-0-165-202]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=24|ip-26-0-171-62]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=9|ip-26-0-166-15]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=17|ip-26-0-172-73]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=17|ip-26-0-171-56]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=30|ip-26-0-173-121]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=9|ip-26-0-171-249]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=27|ip-26-0-173-121]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=21|ip-26-0-163-236]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 9500.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=23|ip-26-0-169-86]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=0|TP=14|ip-26-0-163-226]: [After model building] Memory usage: 28.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=13|ip-26-0-169-207]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=20|ip-26-0-169-86]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=23|ip-26-0-169-86]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=12|ip-26-0-172-57]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=11|ip-26-0-169-207]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=25|ip-26-0-175-165]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=30|ip-26-0-173-121]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=17|ip-26-0-175-132]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=13|ip-26-0-169-207]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=21|ip-26-0-169-86]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=9|ip-26-0-165-131]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=20|ip-26-0-169-86]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=22|ip-26-0-169-86]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=12|ip-26-0-172-57]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=25|ip-26-0-175-165]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=11|ip-26-0-169-207]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=23|ip-26-0-172-73]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=17|ip-26-0-175-132]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 9494.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=21|ip-26-0-169-86]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=9|ip-26-0-165-131]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=20|ip-26-0-172-73]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=11|ip-26-0-166-15]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=22|ip-26-0-166-214]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=11|ip-26-0-172-57]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=12|ip-26-0-174-36]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=12|ip-26-0-171-249]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=21|ip-26-0-172-73]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=23|ip-26-0-172-73]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=23|ip-26-0-166-214]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=22|ip-26-0-169-86]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=28|ip-26-0-170-143]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=11|ip-26-0-166-15]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=20|ip-26-0-172-73]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=22|ip-26-0-166-214]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=11|ip-26-0-172-57]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=12|ip-26-0-174-36]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=12|ip-26-0-171-249]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=21|ip-26-0-172-73]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=28|ip-26-0-170-143]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=23|ip-26-0-165-164]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=23|ip-26-0-171-56]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=18|ip-26-0-169-86]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=18|ip-26-0-172-73]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=23|ip-26-0-166-214]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=22|ip-26-0-165-164]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=20|ip-26-0-175-132]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=11|ip-26-0-165-131]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=23|ip-26-0-165-164]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=11|ip-26-0-174-36]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=23|ip-26-0-171-56]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=26|ip-26-0-170-143]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=18|ip-26-0-172-73]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=28|ip-26-0-173-121]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=21|ip-26-0-175-132]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=18|ip-26-0-169-86]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=22|ip-26-0-165-164]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=28|ip-26-0-166-244]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=10|ip-26-0-169-207]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=20|ip-26-0-175-132]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 9494.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=24|ip-26-0-175-165]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=30|ip-26-0-168-238]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=11|ip-26-0-165-131]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=27|ip-26-0-171-62]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=18|ip-26-0-171-56]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=26|ip-26-0-170-143]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=11|ip-26-0-174-36]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=28|ip-26-0-173-121]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=22|ip-26-0-175-132]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=20|ip-26-0-166-214]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=28|ip-26-0-166-244]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=10|ip-26-0-169-207]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=13|ip-26-0-166-15]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=24|ip-26-0-175-165]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=13|ip-26-0-172-57]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=21|ip-26-0-175-132]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 9494.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=22|ip-26-0-172-73]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=30|ip-26-0-168-238]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=27|ip-26-0-171-62]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=29|ip-26-0-170-143]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=18|ip-26-0-171-56]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=31|ip-26-0-170-143]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=21|ip-26-0-166-214]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=20|ip-26-0-166-214]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=13|ip-26-0-166-15]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=22|ip-26-0-175-132]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 9494.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=8|ip-26-0-171-249]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=13|ip-26-0-172-57]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=8|ip-26-0-174-36]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=29|ip-26-0-170-143]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=28|ip-26-0-175-165]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=22|ip-26-0-172-73]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=17|ip-26-0-168-120]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=12|ip-26-0-165-131]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=27|ip-26-0-168-238]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=10|ip-26-0-166-15]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=21|ip-26-0-166-214]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=8|ip-26-0-165-131]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=8|ip-26-0-171-249]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=26|ip-26-0-166-244]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=31|ip-26-0-170-143]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=8|ip-26-0-174-36]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=31|ip-26-0-173-121]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=28|ip-26-0-175-165]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=13|ip-26-0-171-249]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=13|ip-26-0-167-9]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=12|ip-26-0-165-131]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=17|ip-26-0-168-120]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=13|ip-26-0-174-36]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=20|ip-26-0-168-120]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=28|ip-26-0-165-202]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=10|ip-26-0-166-15]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=27|ip-26-0-168-238]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=13|ip-26-0-165-131]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=29|ip-26-0-171-62]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=29|ip-26-0-173-121]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=21|ip-26-0-165-164]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=13|ip-26-0-167-9]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=21|ip-26-0-168-120]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=31|ip-26-0-173-121]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=26|ip-26-0-166-244]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=13|ip-26-0-171-249]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=30|ip-26-0-171-62]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=9|ip-26-0-167-9]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=13|ip-26-0-174-36]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=20|ip-26-0-168-120]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=10|ip-26-0-165-131]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=8|ip-26-0-165-131]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=13|ip-26-0-165-131]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=18|ip-26-0-166-214]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=9|ip-26-0-174-36]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=10|ip-26-0-167-9]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=29|ip-26-0-173-121]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=21|ip-26-0-165-164]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=28|ip-26-0-165-202]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=29|ip-26-0-171-62]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=21|ip-26-0-168-120]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=9|ip-26-0-167-9]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=30|ip-26-0-171-62]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=10|ip-26-0-165-131]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=29|ip-26-0-175-165]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=18|ip-26-0-166-214]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=9|ip-26-0-174-36]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=10|ip-26-0-167-9]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=24|ip-26-0-165-202]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=30|ip-26-0-175-165]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=29|ip-26-0-175-165]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=18|ip-26-0-168-120]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=24|ip-26-0-165-202]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=27|ip-26-0-175-165]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=31|ip-26-0-165-202]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=10|ip-26-0-172-57]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=31|ip-26-0-166-244]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=11|ip-26-0-171-249]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=14|ip-26-0-169-207]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=30|ip-26-0-175-165]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=14|ip-26-0-172-57]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=26|ip-26-0-173-121]: Local number of parameters: 8.39M (16.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=18|ip-26-0-168-120]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=29|ip-26-0-166-244]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=27|ip-26-0-175-165]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=31|ip-26-0-166-244]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=31|ip-26-0-165-202]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=14|ip-26-0-172-57]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=4|TP=14|ip-26-0-169-207]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=11|ip-26-0-171-249]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=10|ip-26-0-172-57]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=6|TP=26|ip-26-0-173-121]: [After model building] Memory usage: 16.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=29|ip-26-0-166-244]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=26|ip-26-0-171-62]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=10|ip-26-0-174-36]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=26|ip-26-0-175-165]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=10|ip-26-0-171-249]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=14|ip-26-0-174-36]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=14|ip-26-0-166-15]: Local number of parameters: 4.2M (8.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=18|ip-26-0-175-132]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=28|ip-26-0-171-62]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=26|ip-26-0-171-62]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=10|ip-26-0-174-36]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=10|ip-26-0-171-249]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=20|ip-26-0-165-164]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=26|ip-26-0-175-165]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=22|ip-26-0-168-120]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=14|ip-26-0-174-36]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=23|ip-26-0-168-120]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=11|ip-26-0-167-9]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=2|TP=14|ip-26-0-166-15]: [After model building] Memory usage: 8.03MiB. Peak allocated: 5380.00MiB Peak reserved: 10522.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=23|ip-26-0-175-132]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=28|ip-26-0-171-62]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=18|ip-26-0-175-132]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 9494.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=22|ip-26-0-168-120]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=20|ip-26-0-165-164]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=31|ip-26-0-175-165]: Local number of parameters: 0 (0.00MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=23|ip-26-0-168-120]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=11|ip-26-0-167-9]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=20|ip-26-0-171-56]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=23|ip-26-0-175-132]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 9494.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=14|ip-26-0-167-9]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=7|TP=31|ip-26-0-175-165]: [After model building] Memory usage: 0.01MiB. Peak allocated: 5380.00MiB Peak reserved: 10518.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=21|ip-26-0-171-56]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=14|ip-26-0-167-9]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=20|ip-26-0-171-56]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=21|ip-26-0-171-56]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=26|ip-26-0-165-202]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=28|ip-26-0-168-238]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=22|ip-26-0-171-56]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=26|ip-26-0-165-202]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=14|ip-26-0-171-249]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=28|ip-26-0-168-238]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=22|ip-26-0-171-56]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=14|ip-26-0-171-249]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=31|ip-26-0-171-62]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=5|TP=31|ip-26-0-171-62]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=26|ip-26-0-168-238]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=14|ip-26-0-165-131]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=26|ip-26-0-168-238]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=29|ip-26-0-165-202]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=18|ip-26-0-165-164]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=14|ip-26-0-165-131]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=29|ip-26-0-165-202]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=1|TP=18|ip-26-0-165-164]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=31|ip-26-0-168-238]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=29|ip-26-0-168-238]: Local number of parameters: 6.3M (12.02MiB) +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=31|ip-26-0-168-238]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +12/31/2024 00:39:52 [INFO|DP=0|PP=3|TP=29|ip-26-0-168-238]: [After model building] Memory usage: 12.04MiB. Peak allocated: 5380.00MiB Peak reserved: 10524.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/31/2024 00:39:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: [Optimizer Building] Using LearningRateForSP as learning rate +12/31/2024 00:39:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: [ZeRO sharding] Size of optimizer params per rank: +12/31/2024 00:39:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: [ZeRO sharding] DP Rank 0 has 7.35M out of 14.7M (50.00%) params' optimizer states +12/31/2024 00:39:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: [ZeRO sharding] DP Rank 1 has 7.35M out of 14.7M (50.00%) params' optimizer states +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.token_position_embeddings | PP: 0/8 | Block rank: 0 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.0 | PP: 0/8 | Block rank: 1 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.1 | PP: 0/8 | Block rank: 2 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.2 | PP: 0/8 | Block rank: 3 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.3 | PP: 1/8 | Block rank: 0 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.4 | PP: 1/8 | Block rank: 1 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.5 | PP: 1/8 | Block rank: 2 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.6 | PP: 2/8 | Block rank: 0 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.7 | PP: 2/8 | Block rank: 1 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.8 | PP: 3/8 | Block rank: 0 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.9 | PP: 3/8 | Block rank: 1 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.10 | PP: 3/8 | Block rank: 2 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.11 | PP: 4/8 | Block rank: 0 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.12 | PP: 4/8 | Block rank: 1 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.13 | PP: 5/8 | Block rank: 0 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.14 | PP: 5/8 | Block rank: 1 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.decoder.15 | PP: 5/8 | Block rank: 2 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.final_layer_norm | PP: 6/8 | Block rank: 0 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.lm_head | PP: 6/8 | Block rank: 1 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: model.cast_to_fp32 | PP: 7/8 | Block rank: 0 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: module_name: loss | PP: 7/8 | Block rank: 1 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Using dummy data generator +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: [Training Plan] There are 1 training stages +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: [Stage Stable Training Stage] start from step 1 +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: [Start training] datetime: 2024-12-31 00:39:55.050379 | mbs: 8 | grad_accum: 16 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/31/2024 00:39:55 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Memory usage: 112.11MiB. Peak allocated 5380.00MiB. Peak reserved: 10524.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +12/31/2024 00:40:43 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Memory usage: 433.89MiB. Peak allocated 3128.46MiB. Peak reserved: 3498.00MiB +12/31/2024 00:40:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Memory usage: 489.94MiB. Peak allocated 560.00MiB. Peak reserved: 3500.00MiB +12/31/2024 00:40:54 [INFO|DP=0|PP=7|TP=0|ip-26-0-174-186]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 60.4K | tokens_per_sec: 17.4K | tokens_per_sec_per_gpu: 33.9 | global_batch_size: 256 | lm_loss: 12 | lr: 0.00015 | model_tflops_per_gpu: 0.328 | hardware_tflops_per_gpu: 0.328 | grad_norm: 0.268 | cuda_memory_allocated: 314K | cuda_max_memory_reserved: 1.88G | hd_total_memory_tb: 312G | hd_used_memory_tb: 247G | hd_free_memory_tb: 65.5G +12/31/2024 00:41:20 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Memory usage: 489.94MiB. Peak allocated 3127.45MiB. Peak reserved: 3766.00MiB +12/31/2024 00:41:21 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Memory usage: 489.94MiB. Peak allocated 560.00MiB. Peak reserved: 3766.00MiB +12/31/2024 00:41:21 [INFO|DP=0|PP=7|TP=0|ip-26-0-174-186]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 26.3K | tokens_per_sec: 39.9K | tokens_per_sec_per_gpu: 78 | global_batch_size: 256 | lm_loss: 12 | lr: 0.0003 | model_tflops_per_gpu: 0.754 | hardware_tflops_per_gpu: 0.754 | grad_norm: 0.269 | cuda_memory_allocated: 314K | cuda_max_memory_reserved: 1.88G | hd_total_memory_tb: 312G | hd_used_memory_tb: 247G | hd_free_memory_tb: 65.5G +12/31/2024 00:41:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Memory usage: 489.94MiB. Peak allocated 3127.45MiB. Peak reserved: 3894.00MiB +12/31/2024 00:41:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +12/31/2024 00:41:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: | -------- | --------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | ----------------------------------- | -------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +12/31/2024 00:41:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: | 14016454 | 1.34G_dp2_tp32_pp8_acc16_mbs8_seq4096_zero1_tpmodeRED_vocab131k | 64 | 4096 | 8 | 16 | 256 | 0.76 | 0.76 | 78.75 | 12.94 | 6.55 | 6.32 | 423.36 | 241.04 | 241.22 | 0.55 | 3.80 | 2 | 8 | 32 | OneForwardOneBackwardPipelineEngine | REDUCE_SCATTER | True | 2048 | silu | 16 | 32 | 32 | 4096 | 131072 | True | torch.bfloat16 | 1 | 25 | True | 1.61G | 14.7M | +12/31/2024 00:41:47 [INFO|DP=0|PP=7|TP=0|ip-26-0-174-186]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 26K | tokens_per_sec: 40.3K | tokens_per_sec_per_gpu: 78.8 | global_batch_size: 256 | lm_loss: 12 | lr: 0.000296 | model_tflops_per_gpu: 0.761 | hardware_tflops_per_gpu: 0.761 | grad_norm: 0.238 | cuda_memory_allocated: 314K | cuda_max_memory_reserved: 1.88G | hd_total_memory_tb: 312G | hd_used_memory_tb: 247G | hd_free_memory_tb: 65.5G +12/31/2024 00:41:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final2.csv +12/31/2024 00:41:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Acquired lock for CSV file: benchmark/results/bench_final2.csv +12/31/2024 00:41:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Successfully wrote to CSV file: benchmark/results/bench_final2.csv. Releasing lock... +12/31/2024 00:41:47 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-220]: Throughput logging complete +terminate called after throwing an instance of 'c10::Error' + what(): CUDA driver error: unknown error +Exception raised from _hasPrimaryContext at ../aten/src/ATen/cuda/detail/CUDAHooks.cpp:67 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fc76b026617 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10.so) +frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, char const*) + 0x68 (0x7fc76afe1a56 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10.so) +frame #2: + 0x11a00bf (0x7fc76c3420bf in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #3: c10::cuda::MaybeSetDevice(int) + 0xc (0x7fc76b0d7bdc in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10_cuda.so) +frame #4: std::_Sp_counted_ptr_inplace >, std::allocator > >, (__gnu_cxx::_Lock_policy)2>::_M_dispose() + 0x98 (0x7fc76c451ae8 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #5: std::_Sp_counted_base<(__gnu_cxx::_Lock_policy)2>::_M_release() + 0x48 (0x7fc7ca3a4048 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_python.so) +frame #6: c10d::ProcessGroupNCCL::WorkNCCL::~WorkNCCL() + 0x135 (0x7fc76c41cc55 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #7: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x3c5 (0x7fc76c432585 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #8: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x78 (0x7fc76c432718 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #9: + 0xd3e95 (0x7fc7fa12ae95 in /fsx/nouamane/miniconda/envs/2-1-cu121/bin/../lib/libstdc++.so.6) +frame #10: + 0x8609 (0x7fc7fa59a609 in /lib/x86_64-linux-gnu/libpthread.so.0) +frame #11: clone + 0x43 (0x7fc7fa363353 in /lib/x86_64-linux-gnu/libc.so.6) + +terminate called after throwing an instance of 'c10::Error' + what(): CUDA driver error: unknown error +Exception raised from _hasPrimaryContext at ../aten/src/ATen/cuda/detail/CUDAHooks.cpp:67 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fc93ba64617 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10.so) +frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, char const*) + 0x68 (0x7fc93ba1fa56 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10.so) +frame #2: + 0x11a00bf (0x7fc93cd800bf in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #3: c10::cuda::MaybeSetDevice(int) + 0xc (0x7fc93bb15bdc in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10_cuda.so) +frame #4: std::_Sp_counted_ptr_inplace >, std::allocator > >, (__gnu_cxx::_Lock_policy)2>::_M_dispose() + 0x98 (0x7fc93ce8fae8 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #5: std::_Sp_counted_base<(__gnu_cxx::_Lock_policy)2>::_M_release() + 0x48 (0x7fc99ade2048 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_python.so) +frame #6: c10d::ProcessGroupNCCL::WorkNCCL::~WorkNCCL() + 0x135 (0x7fc93ce5ac55 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #7: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x3c5 (0x7fc93ce70585 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #8: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x78 (0x7fc93ce70718 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #9: + 0xd3e95 (0x7fc9cab68e95 in /fsx/nouamane/miniconda/envs/2-1-cu121/bin/../lib/libstdc++.so.6) +frame #10: + 0x8609 (0x7fc9cafd8609 in /lib/x86_64-linux-gnu/libpthread.so.0) +frame #11: clone + 0x43 (0x7fc9cada1353 in /lib/x86_64-linux-gnu/libc.so.6) + +double free or corruption (!prev) +free(): invalid pointer +free(): invalid pointer +terminate called after throwing an instance of 'c10::Error' + what(): CUDA driver error: unknown error +Exception raised from _hasPrimaryContext at ../aten/src/ATen/cuda/detail/CUDAHooks.cpp:67 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7f157bcc2617 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10.so) +frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, char const*) + 0x68 (0x7f157bc7da56 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10.so) +frame #2: + 0x11a00bf (0x7f157cfde0bf in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #3: c10::cuda::MaybeSetDevice(int) + 0xc (0x7f157bd73bdc in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10_cuda.so) +frame #4: std::_Sp_counted_ptr_inplace >, std::allocator > >, (__gnu_cxx::_Lock_policy)2>::_M_dispose() + 0x98 (0x7f157d0edae8 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #5: std::_Sp_counted_base<(__gnu_cxx::_Lock_policy)2>::_M_release() + 0x48 (0x7f15db040048 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_python.so) +frame #6: c10d::ProcessGroupNCCL::WorkNCCL::~WorkNCCL() + 0x135 (0x7f157d0b8c55 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #7: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x3c5 (0x7f157d0ce585 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #8: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x78 (0x7f157d0ce718 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #9: + 0xd3e95 (0x7f160adc6e95 in /fsx/nouamane/miniconda/envs/2-1-cu121/bin/../lib/libstdc++.so.6) +frame #10: + 0x8609 (0x7f160b236609 in /lib/x86_64-linux-gnu/libpthread.so.0) +frame #11: clone + 0x43 (0x7f160afff353 in /lib/x86_64-linux-gnu/libc.so.6) + +terminate called after throwing an instance of 'c10::Error' + what(): CUDA driver error: unknown error +Exception raised from _hasPrimaryContext at ../aten/src/ATen/cuda/detail/CUDAHooks.cpp:67 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fe7faa14617 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10.so) +frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, char const*) + 0x68 (0x7fe7fa9cfa56 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10.so) +frame #2: + 0x11a00bf (0x7fe7fbd300bf in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #3: c10::cuda::MaybeSetDevice(int) + 0xc (0x7fe7faac5bdc in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10_cuda.so) +frame #4: std::_Sp_counted_ptr_inplace >, std::allocator > >, (__gnu_cxx::_Lock_policy)2>::_M_dispose() + 0x98 (0x7fe7fbe3fae8 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #5: std::_Sp_counted_base<(__gnu_cxx::_Lock_policy)2>::_M_release() + 0x48 (0x7fe859d92048 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_python.so) +frame #6: c10d::ProcessGroupNCCL::WorkNCCL::~WorkNCCL() + 0x135 (0x7fe7fbe0ac55 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #7: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x3c5 (0x7fe7fbe20585 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #8: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x78 (0x7fe7fbe20718 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #9: + 0xd3e95 (0x7fe889b18e95 in /fsx/nouamane/miniconda/envs/2-1-cu121/bin/../lib/libstdc++.so.6) +frame #10: + 0x8609 (0x7fe889f88609 in /lib/x86_64-linux-gnu/libpthread.so.0) +frame #11: clone + 0x43 (0x7fe889d51353 in /lib/x86_64-linux-gnu/libc.so.6) + +[2024-12-31 00:42:14,341] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 69870 closing signal SIGTERM +[2024-12-31 00:42:14,341] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 69871 closing signal SIGTERM +[2024-12-31 00:42:14,341] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 69872 closing signal SIGTERM +[2024-12-31 00:42:14,341] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 69874 closing signal SIGTERM +[2024-12-31 00:42:14,341] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 69875 closing signal SIGTERM +[2024-12-31 00:42:14,341] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 69876 closing signal SIGTERM +[2024-12-31 00:42:14,341] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 69877 closing signal SIGTERM +[2024-12-31 00:42:17,573] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 3 (pid: 69873) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2024-12-31_00:42:14 + host : ip-26-0-175-34.ec2.internal + rank : 507 (local_rank: 3) + exitcode : 1 (pid: 69873) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-175-34: task 59: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=14016454.0 +slurmstepd: error: *** STEP 14016454.0 ON ip-26-0-163-220 CANCELLED AT 2024-12-31T00:42:17 *** +[2024-12-31 00:42:17,872] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,871] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,871] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 324025 closing signal SIGTERM +[2024-12-31 00:42:17,871] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 324026 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 324028 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 333239 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 324029 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 324030 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 333240 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 614774 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 333241 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 324031 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 614775 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 333242 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 324032 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 333243 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 614776 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 333244 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 614777 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 614778 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83694 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 614779 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83695 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 614780 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 614781 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83696 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83697 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,872] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83698 closing signal SIGTERM +[2024-12-31 00:42:17,871] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2992402 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83699 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 43498 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 151485 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 502666 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 133264 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83700 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 151486 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 502667 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 319981 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83701 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2992404 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 43499 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 151487 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 680493 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 502668 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 319982 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 43500 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 151488 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 502669 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 133265 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271265 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 319983 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2992405 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 43501 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 151489 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 680494 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 502670 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271266 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 319984 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 43502 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 680495 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 502671 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 133266 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271267 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 319985 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2992406 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 97596 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 43503 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 680496 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271268 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2992407 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 97597 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 151490 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 680497 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271269 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 319986 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2992408 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 97598 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 151491 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 502672 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 133267 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 319987 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 97599 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 43504 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 151492 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 502673 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236985 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 319988 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2992409 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 97600 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 43505 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 680498 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271270 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 97601 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 680499 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 133268 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 134946 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236986 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271271 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106973 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 680500 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 133269 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 134947 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236987 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271272 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 97602 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106974 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 133270 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 134948 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26290 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236988 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 97603 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 167928 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106975 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 134949 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 167929 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106976 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 134950 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26291 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 521257 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106977 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 133271 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 134951 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 167930 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96181 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106978 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26292 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236989 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 521258 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106979 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236990 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 378350 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338798 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96182 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 134952 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236991 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 167931 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 134953 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26293 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236992 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174669 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 378351 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338799 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 521259 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96183 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106980 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26294 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174670 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 378352 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 167932 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96184 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 378353 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74771 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338800 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 167933 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174671 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338801 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 521260 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236830 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26295 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174672 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82420 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74772 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236831 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96185 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26296 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 378354 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82421 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74773 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 167934 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96186 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 26297 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 378355 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82422 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74774 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338802 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 167935 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 521261 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236832 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96187 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 368371 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174673 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 378356 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82423 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74775 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338803 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 521262 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236833 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96188 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 368372 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174674 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 378357 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82424 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338804 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 368373 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174675 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82425 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338805 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 368374 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74776 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 521263 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236834 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 368375 closing signal SIGTERM +[2024-12-31 00:42:17,872] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174676 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82426 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74777 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 521264 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 352668 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82427 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74778 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236835 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 963310 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 352669 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 368376 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61336 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236836 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 963311 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 368377 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61337 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 236837 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 352670 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 107580 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 368378 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 963312 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 107581 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104220 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61338 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 107582 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 295612 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 352671 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 107583 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104221 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 963313 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 352672 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 107584 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37702 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 963314 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 352673 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 107585 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61339 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 295613 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 107586 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37703 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104222 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61340 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37704 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140150 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61341 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 295614 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 963315 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 352674 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 223340 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37705 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61342 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 295615 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 963316 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 352675 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 107587 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37706 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104223 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 963317 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 223341 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140151 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 223342 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61343 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 295616 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100405 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 223343 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104224 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 295617 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100406 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 223344 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37707 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140152 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 295618 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 408583 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100407 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37708 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104225 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100408 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37709 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 408584 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100409 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 223345 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140153 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 118580 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 295619 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100410 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 223346 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104226 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 408585 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 369286 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 223347 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 118581 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140154 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100411 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 369287 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140155 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104227 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 408586 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68921 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 100412 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 118582 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82649 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 408587 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68922 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 369288 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 408588 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68923 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 369289 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140156 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82651 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68924 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 341756 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 140157 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 118583 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68925 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82652 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 408589 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68926 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 341757 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196252 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 369290 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82653 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 408590 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68927 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 369291 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 118584 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 251701 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 341758 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196029 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 337420 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196253 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 369292 closing signal SIGTERM +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3184716 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 118585 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82654 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196254 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,873] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3184717 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82655 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 251702 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68928 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 341759 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196030 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 337421 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196255 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3184718 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 118586 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 251703 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271552 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 341760 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196031 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 369293 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2024954 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3184719 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 118587 closing signal SIGTERM +[2024-12-31 00:42:17,877] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 251704 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 341761 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196032 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2024955 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3184720 closing signal SIGTERM +[2024-12-31 00:42:17,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 251705 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271553 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 341762 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 337422 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196256 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2024956 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 73063 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3184721 closing signal SIGTERM +[2024-12-31 00:42:17,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91228 closing signal SIGTERM +[2024-12-31 00:42:17,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 251706 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271554 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196257 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2024957 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91229 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271555 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 341763 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196033 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196258 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2024958 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91230 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196034 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 337423 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196259 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2024959 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3184722 closing signal SIGTERM +[2024-12-31 00:42:17,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91231 closing signal SIGTERM +[2024-12-31 00:42:17,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 251707 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196035 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 337424 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90761 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 73064 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103613 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3184723 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91232 closing signal SIGTERM +[2024-12-31 00:42:17,877] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 251708 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271556 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 196036 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271557 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2024960 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90762 closing signal SIGTERM +[2024-12-31 00:42:17,878] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 73065 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103614 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271558 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 337425 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2024961 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90763 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114479 closing signal SIGTERM +[2024-12-31 00:42:17,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91233 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 337426 closing signal SIGTERM +[2024-12-31 00:42:17,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68718 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103615 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114480 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96148 closing signal SIGTERM +[2024-12-31 00:42:17,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91234 closing signal SIGTERM +[2024-12-31 00:42:17,874] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2573446 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 73066 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103616 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114481 closing signal SIGTERM +[2024-12-31 00:42:17,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91235 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 271559 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90764 closing signal SIGTERM +[2024-12-31 00:42:17,878] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68719 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 243757 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96149 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 337427 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90765 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 243758 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114482 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2573447 closing signal SIGTERM +[2024-12-31 00:42:17,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68720 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 73067 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103617 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96150 closing signal SIGTERM +[2024-12-31 00:42:17,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68721 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 243760 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103618 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96151 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90766 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103619 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114483 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108429 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2573448 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90767 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 73068 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 103620 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108430 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71073 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174641 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90768 closing signal SIGTERM +[2024-12-31 00:42:17,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68722 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 243761 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96152 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108431 closing signal SIGTERM +[2024-12-31 00:42:17,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68723 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 243762 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114484 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96153 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108432 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2573449 closing signal SIGTERM +[2024-12-31 00:42:17,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68724 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 73069 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114485 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96154 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108433 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71074 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174642 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 73070 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 96155 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108434 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174643 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 243763 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71075 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2573450 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174644 closing signal SIGTERM +[2024-12-31 00:42:17,879] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 68725 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 243764 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 114486 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174645 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108435 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108436 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71076 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71077 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2573451 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2573452 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174646 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174647 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71078 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 174648 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71079 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2573453 closing signal SIGTERM +[2024-12-31 00:42:17,875] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72572 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72573 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72574 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72575 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72576 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72577 closing signal SIGTERM +[2024-12-31 00:42:17,876] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 72578 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 333146 got signal: 15 +srun: error: ip-26-0-163-220: task 0: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 72496 got signal: 15 +srun: error: ip-26-0-175-165: task 61: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2992328 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 82559 got signal: 15 +srun: error: ip-26-0-164-18: task 4: Exited with exit code 1 +srun: error: ip-26-0-174-186: task 57: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 70997 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 323950 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 368296 got signal: 15 +srun: error: ip-26-0-175-170: task 62: Exited with exit code 1 +[2024-12-31 00:42:23,340] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-175-132.ec2.internal_108340_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-164-0: task 3: Exited with exit code 1 +srun: error: ip-26-0-166-244: task 20: Exited with exit code 1 +[2024-12-31 00:42:23,448] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-214.ec2.internal_236910_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,461] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-38.ec2.internal_82346_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,463] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-173-7.ec2.internal_90678_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,475] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-120.ec2.internal_378275_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,485] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-51.ec2.internal_338724_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,491] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-36.ec2.internal_106898_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,496] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-45.ec2.internal_43421_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,508] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-174-100.ec2.internal_114395_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,511] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-238.ec2.internal_107504_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,513] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-9.ec2.internal_521182_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,537] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-142.ec2.internal_103538_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,548] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-59.ec2.internal_134872_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,551] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-213.ec2.internal_26211_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,554] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-88.ec2.internal_223251_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,559] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-52.ec2.internal_61261_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,563] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-168.ec2.internal_196176_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,573] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-147.ec2.internal_243681_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,583] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-95.ec2.internal_37627_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,595] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-163-236.ec2.internal_133190_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,599] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-62.ec2.internal_295528_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,607] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-249.ec2.internal_337344_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,609] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-207.ec2.internal_167851_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,614] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-236.ec2.internal_502589_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,615] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-125.ec2.internal_680403_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,625] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-174-36.ec2.internal_174566_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,626] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-173-121.ec2.internal_2573372_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 134872 got signal: 15 +[2024-12-31 00:42:23,648] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-102.ec2.internal_68834_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,650] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-56.ec2.internal_118499_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,653] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-139.ec2.internal_271477_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,655] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-24.ec2.internal_319892_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,664] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-170-160.ec2.internal_100329_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,667] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-131.ec2.internal_151395_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,669] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-163-226.ec2.internal_614700_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,672] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-187.ec2.internal_83620_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,684] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-164.ec2.internal_271189_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,701] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-15.ec2.internal_174594_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,703] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-207.ec2.internal_104145_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,706] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-202.ec2.internal_97521_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,710] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-230.ec2.internal_91152_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,713] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-116.ec2.internal_72989_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,713] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-30.ec2.internal_352593_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,718] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-57.ec2.internal_251626_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,724] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-21.ec2.internal_140067_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,724] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-173-246.ec2.internal_3184627_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,730] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-34.ec2.internal_236753_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 243681 got signal: 15 +[2024-12-31 00:42:23,737] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-86.ec2.internal_369211_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,741] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-170-143.ec2.internal_408506_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,776] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-252.ec2.internal_195953_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:23,850] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-173-202.ec2.internal_2024879_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-165-59: task 11: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 108340 got signal: 15 +[2024-12-31 00:42:23,977] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-73.ec2.internal_341681_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-172-147: task 49: Exited with exit code 1 +[2024-12-31 00:42:24,046] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-175.ec2.internal_96106_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:24,057] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-175-241.ec2.internal_68644_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 37627 got signal: 15 +[2024-12-31 00:42:24,090] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-177.ec2.internal_74696_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:24,191] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-175-19.ec2.internal_96072_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 97521 got signal: 15 + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 174566 got signal: 15 +[2024-12-31 00:42:24,239] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-217.ec2.internal_963235_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-175-132: task 60: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 26211 got signal: 15 +srun: error: ip-26-0-168-95: task 29: Exited with exit code 1 +srun: error: ip-26-0-165-202: task 14: Exited with exit code 1 +srun: error: ip-26-0-174-36: task 55: Exited with exit code 1 +srun: error: ip-26-0-165-213: task 15: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 680403 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 103538 got signal: 15 +srun: error: ip-26-0-166-125: task 18: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 43421 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 96072 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 106898 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 140067 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 68644 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 408506 got signal: 15 + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 91152 got signal: 15 +srun: error: ip-26-0-172-142: task 48: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 223251 got signal: 15 +srun: error: ip-26-0-166-36: task 17: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler +srun: error: ip-26-0-164-45: task 5: Exited with exit code 1 + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 72989 got signal: 15 +srun: error: ip-26-0-171-21: task 37: Exited with exit code 1 +srun: error: ip-26-0-175-19: task 58: Exited with exit code 1 +srun: error: ip-26-0-175-241: task 63: Exited with exit code 1 +srun: error: ip-26-0-171-88: task 40: Exited with exit code 1 +srun: error: ip-26-0-170-143: task 35: Exited with exit code 1 +srun: error: ip-26-0-171-230: task 43: Exited with exit code 1 +srun: error: ip-26-0-172-116: task 47: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2573372 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 502589 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 100329 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 68834 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 338724 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 107504 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 295528 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 352593 got signal: 15 +srun: error: ip-26-0-164-236: task 8: Exited with exit code 1 +srun: error: ip-26-0-173-121: task 52: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 271477 got signal: 15 +srun: error: ip-26-0-170-160: task 36: Exited with exit code 1 +srun: error: ip-26-0-171-102: task 41: Exited with exit code 1 +srun: error: ip-26-0-168-238: task 31: Exited with exit code 1 +srun: error: ip-26-0-167-51: task 22: Exited with exit code 1 +srun: error: ip-26-0-168-30: task 26: Exited with exit code 1 +srun: error: ip-26-0-171-62: task 39: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 133190 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 319892 got signal: 15 +srun: error: ip-26-0-169-139: task 33: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 83620 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 151395 got signal: 15 +srun: error: ip-26-0-165-24: task 9: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 74696 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 174594 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 167851 got signal: 15 +srun: error: ip-26-0-164-187: task 6: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + sys.exit(main()) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + elastic_launch( + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 61261 got signal: 15 + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler +srun: error: ip-26-0-163-236: task 2: Exited with exit code 1 + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 963235 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 104145 got signal: 15 +srun: error: ip-26-0-165-131: task 12: Exited with exit code 1 +srun: error: ip-26-0-167-177: task 24: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 337344 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 378275 got signal: 15 +srun: error: ip-26-0-164-207: task 7: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 521182 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 236753 got signal: 15 +srun: error: ip-26-0-166-15: task 16: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run +srun: error: ip-26-0-169-207: task 34: Exited with exit code 1 + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 614700 got signal: 15 +srun: error: ip-26-0-167-217: task 25: Exited with exit code 1 +srun: error: ip-26-0-168-52: task 28: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 196176 got signal: 15 +srun: error: ip-26-0-168-34: task 27: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run +srun: error: ip-26-0-171-249: task 44: Exited with exit code 1 + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 96106 got signal: 15 +srun: error: ip-26-0-168-120: task 30: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 82346 got signal: 15 +srun: error: ip-26-0-163-226: task 1: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run +srun: error: ip-26-0-167-9: task 21: Exited with exit code 1 + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 369211 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 271189 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 195953 got signal: 15 +srun: error: ip-26-0-171-168: task 42: Exited with exit code 1 +srun: error: ip-26-0-167-175: task 23: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 341681 got signal: 15 +srun: error: ip-26-0-169-86: task 32: Exited with exit code 1 +srun: error: ip-26-0-165-164: task 13: Exited with exit code 1 +srun: error: ip-26-0-172-252: task 50: Exited with exit code 1 +srun: error: ip-26-0-165-38: task 10: Exited with exit code 1 +srun: error: ip-26-0-172-73: task 46: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 114395 got signal: 15 +srun: error: ip-26-0-174-100: task 56: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 236910 got signal: 15 +[2024-12-31 00:42:28,465] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-173-7.ec2.internal_90678_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:28,652] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-171-56.ec2.internal_118499_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 90678 got signal: 15 +srun: error: ip-26-0-166-214: task 19: Exited with exit code 1 +[2024-12-31 00:42:28,719] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-172-57.ec2.internal_251626_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:28,726] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-173-246.ec2.internal_3184627_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +[2024-12-31 00:42:28,851] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-173-202.ec2.internal_2024879_0' has failed to send a keep-alive heartbeat to the rendezvous '14016454' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2024879 got signal: 15 +srun: error: ip-26-0-173-7: task 51: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 118499 got signal: 15 +srun: error: ip-26-0-173-202: task 53: Exited with exit code 1 +srun: error: ip-26-0-171-56: task 38: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 3184627 got signal: 15 +srun: error: ip-26-0-173-246: task 54: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 251626 got signal: 15 +srun: error: ip-26-0-172-57: task 45: Exited with exit code 1 diff --git a/logs/14018457-bench_1.34G_dp16_tp1_pp2_acc16_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/14018457-bench_1.34G_dp16_tp1_pp2_acc16_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..f91391ebf887851d9f76e8c99643cc9a1b985a8e --- /dev/null +++ b/logs/14018457-bench_1.34G_dp16_tp1_pp2_acc16_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,1608 @@ ++ '[' -z 14018457 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-163-43,ip-26-0-170-160,ip-26-0-171-[21,56]' ++ export 'NODELIST=ip-26-0-163-43 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56' ++ NODELIST='ip-26-0-163-43 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-163-43,ip-26-0-170-160,ip-26-0-171-[21,56]' ++ export MASTER_NODE=ip-26-0-163-43 ++ MASTER_NODE=ip-26-0-163-43 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-43 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 52-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 52-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 52-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 52-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-163-43' +Master node: ip-26-0-163-43 ++ echo 'All nodes: ip-26-0-163-43 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56' +All nodes: ip-26-0-163-43 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 ++ echo 'World size: 32' +World size: 32 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=14018457 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-163-43:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_1.34G_dp16_tp1_pp2_acc16_mbs1_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2025-01-02 00:40:20,409] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 00:40:20,411] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 00:40:20,412] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 00:40:20,411] torch.distributed.run: [WARNING] +[2025-01-02 00:40:20,411] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 00:40:20,411] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 00:40:20,411] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 00:40:20,409] torch.distributed.run: [WARNING] +[2025-01-02 00:40:20,409] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 00:40:20,409] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 00:40:20,409] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 00:40:20,413] torch.distributed.run: [WARNING] +[2025-01-02 00:40:20,413] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 00:40:20,413] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 00:40:20,413] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 00:40:20,527] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 00:40:20,527] torch.distributed.run: [WARNING] +[2025-01-02 00:40:20,527] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 00:40:20,527] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 00:40:20,527] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Config: +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Config(general=GeneralArgs(project='debug', +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: run='1.34G_dp16_tp1_pp2_acc16_mbs1_seq4096_zero0_tpmodeRED_vocab131k', +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: seed=42, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: step=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: consumed_train_samples=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: ignore_sanity_checks=True), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: parallelism=ParallelismArgs(dp=16, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: pp=2, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: tp=1, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: pp_engine=, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: tp_mode=, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: tp_linear_async_communication=True, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: recompute_layer=False, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: tp_recompute_allgather=True, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: expert_parallel_size=1), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: eos_token_id=0, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: hidden_act='silu', +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: hidden_size=2048, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: initializer_range=0.02, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: intermediate_size=8192, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: is_llama_config=True, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: max_position_embeddings=4096, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: num_attention_heads=32, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: num_hidden_layers=16, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: num_key_value_heads=32, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: pad_token_id=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: pretraining_tp=1, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: rms_norm_eps=1e-05, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: rope_scaling=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: rope_theta=10000.0, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: rope_interleaved=False, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: tie_word_embeddings=True, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: use_cache=True, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: vocab_size=131072), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: init_method=RandomInit(std=0.02), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: dtype=torch.bfloat16, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: make_vocab_size_divisible_by=1, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: ddp_bucket_cap_mb=25), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: tokenizer_revision=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: tokenizer_max_length=None), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: checkpoint_interval=10000, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: save_initial_state=False, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: save_final_state=False, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: resume_checkpoint_path=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: checkpoints_path_is_shared_file_system=False), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: logging=LoggingArgs(log_level='info', +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: log_level_replica='info', +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: iteration_step_info_interval=1), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: tokens=TokensArgs(sequence_length=4096, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: train_steps=100, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: micro_batch_size=1, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: batch_accumulation_per_replica=16, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: val_check_interval=100, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: limit_val_batches=0, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: limit_test_batches=0), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: adam_beta1=0.9, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: adam_beta2=0.95, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: torch_adam_is_fused=True, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: name='adamW'), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: zero_stage=0, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: weight_decay=0.01, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: clip_grad=1.0, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: accumulate_grad_in_fp32=True, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: lr_warmup_steps=2, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: lr_warmup_style='linear', +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: lr_decay_style='cosine', +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: lr_decay_steps=13, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: lr_decay_starting_step=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: min_decay_lr=1e-05)), +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: start_training_step=1, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: data=DataArgs(dataset=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: seed=42, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: num_loading_workers=1))], +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: profiler=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: lighteval=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: s3_upload=None) +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Model Config: +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: LlamaConfig(bos_token_id=0, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: eos_token_id=0, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: hidden_act='silu', +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: hidden_size=2048, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: initializer_range=0.02, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: intermediate_size=8192, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: is_llama_config=True, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: max_position_embeddings=4096, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: num_attention_heads=32, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: num_hidden_layers=16, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: num_key_value_heads=32, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: pad_token_id=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: pretraining_tp=1, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: rms_norm_eps=1e-05, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: rope_scaling=None, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: rope_theta=10000.0, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: rope_interleaved=False, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: tie_word_embeddings=True, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: use_cache=True, +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: vocab_size=131072) +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Building model.. +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Initialize RoPE Theta = 10000.0 +01/02/2025 00:41:03 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Setting PP block ranks... +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/02/2025 00:41:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Total number of parameters: 1.61G (3072.13MiB) +01/02/2025 00:41:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Local number of parameters: 1.01G (1920.09MiB) +01/02/2025 00:41:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: [After model building] Memory usage: 1920.11MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +01/02/2025 00:41:05 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-21]: Local number of parameters: 604M (1152.04MiB) +01/02/2025 00:41:05 [INFO|DP=0|PP=1|TP=0|ip-26-0-171-21]: [After model building] Memory usage: 1152.06MiB. Peak allocated: 5440.00MiB Peak reserved: 14658.00MiB +01/02/2025 00:41:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: No checkpoint path provided. +01/02/2025 00:41:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Parametrizing model parameters using StandardParametrizator +01/02/2025 00:41:13 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: [Optimizer Building] Using LearningRateForSP as learning rate +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.token_position_embeddings | PP: 0/2 | Block rank: 0 + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.0 | PP: 0/2 | Block rank: 1 +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.1 | PP: 0/2 | Block rank: 2 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.2 | PP: 0/2 | Block rank: 3 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.3 | PP: 0/2 | Block rank: 4 + trainer.train(dataloader) +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.4 | PP: 0/2 | Block rank: 5 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.5 | PP: 0/2 | Block rank: 6 + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.6 | PP: 0/2 | Block rank: 7 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.7 | PP: 0/2 | Block rank: 8 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.8 | PP: 0/2 | Block rank: 9 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.9 | PP: 0/2 | Block rank: 10 + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.10 | PP: 0/2 | Block rank: 11 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.11 | PP: 1/2 | Block rank: 0 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.12 | PP: 1/2 | Block rank: 1 + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.13 | PP: 1/2 | Block rank: 2 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.14 | PP: 1/2 | Block rank: 3 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.decoder.15 | PP: 1/2 | Block rank: 4 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.final_layer_norm | PP: 1/2 | Block rank: 5 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.lm_head | PP: 1/2 | Block rank: 6 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: model.cast_to_fp32 | PP: 1/2 | Block rank: 7 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: module_name: loss | PP: 1/2 | Block rank: 8 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples + return f(*args, **kwargs) +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Using dummy data generator + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: [Training Plan] There are 1 training stages +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: [Stage Stable Training Stage] start from step 1 +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: [Start training] datetime: 2025-01-02 00:41:14.550855 | mbs: 1 | grad_accum: 16 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + model.p2p.clear_history()outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +Traceback (most recent call last): + model.p2p.clear_history() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +01/02/2025 00:41:14 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-43]: Memory usage: 11520.54MiB. Peak allocated 11520.54MiB. Peak reserved: 24264.00MiB +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 253, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") +AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' +[2025-01-02 00:41:37,735] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 97258) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2025-01-02 00:41:37,735] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 25461) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2025-01-02 00:41:37,734] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46376 closing signal SIGTERM +[2025-01-02 00:41:37,734] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46377 closing signal SIGTERM +[2025-01-02 00:41:37,734] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46378 closing signal SIGTERM +[2025-01-02 00:41:37,734] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46379 closing signal SIGTERM +[2025-01-02 00:41:37,734] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46380 closing signal SIGTERM +[2025-01-02 00:41:37,734] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46381 closing signal SIGTERM +[2025-01-02 00:41:37,735] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46382 closing signal SIGTERM +[2025-01-02 00:41:37,740] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125696 closing signal SIGTERM +[2025-01-02 00:41:37,740] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125697 closing signal SIGTERM +[2025-01-02 00:41:37,740] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125698 closing signal SIGTERM +[2025-01-02 00:41:37,740] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125699 closing signal SIGTERM +[2025-01-02 00:41:37,740] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125701 closing signal SIGTERM +[2025-01-02 00:41:37,740] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125702 closing signal SIGTERM +[2025-01-02 00:41:37,781] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_pnhpd7t2/14018457_2rewdjir/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main +[2025-01-02 00:41:37,782] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_uzuzl5wj/14018457_7qh_8ojw/attempt_0/0/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2025-01-02_00:41:14 + host : ip-26-0-171-56.ec2.internal + rank : 25 (local_rank: 1) + exitcode : 1 (pid: 25462) + error_file: /tmp/torchelastic_pnhpd7t2/14018457_2rewdjir/attempt_0/1/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[2]: + time : 2025-01-02_00:41:14 + host : ip-26-0-171-56.ec2.internal + rank : 26 (local_rank: 2) + exitcode : 1 (pid: 25463) + error_file: /tmp/torchelastic_pnhpd7t2/14018457_2rewdjir/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[3]: + time : 2025-01-02_00:41:14 + host : ip-26-0-171-56.ec2.internal + rank : 27 (local_rank: 3) + exitcode : 1 (pid: 25464) + error_file: /tmp/torchelastic_pnhpd7t2/14018457_2rewdjir/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[4]: + time : 2025-01-02_00:41:14 + host : ip-26-0-171-56.ec2.internal + rank : 28 (local_rank: 4) + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + exitcode : 1 (pid: 25465) + error_file: /tmp/torchelastic_pnhpd7t2/14018457_2rewdjir/attempt_0/4/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[5]: + time : 2025-01-02_00:41:14 + host : ip-26-0-171-56.ec2.internal + rank : 29 (local_rank: 5) + exitcode : 1 (pid: 25466) + error_file: /tmp/torchelastic_pnhpd7t2/14018457_2rewdjir/attempt_0/5/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[6]: + time : 2025-01-02_00:41:14 + host : ip-26-0-171-56.ec2.internal + rank : 30 (local_rank: 6) + exitcode : 1 (pid: 25467) + error_file: /tmp/torchelastic_pnhpd7t2/14018457_2rewdjir/attempt_0/6/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[7]: + time : 2025-01-02_00:41:14 + host : ip-26-0-171-56.ec2.internal + rank : 31 (local_rank: 7) + exitcode : 1 (pid: 25468) + error_file: /tmp/torchelastic_pnhpd7t2/14018457_2rewdjir/attempt_0/7/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2025-01-02_00:41:14 + host : ip-26-0-171-56.ec2.internal + rank : 24 (local_rank: 0) + exitcode : 1 (pid: 25461) + error_file: /tmp/torchelastic_pnhpd7t2/14018457_2rewdjir/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +============================================================ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2025-01-02_00:41:14 + host : ip-26-0-163-43.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 97259) + error_file: /tmp/torchelastic_uzuzl5wj/14018457_7qh_8ojw/attempt_0/1/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[2]: + time : 2025-01-02_00:41:14 + host : ip-26-0-163-43.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 97260) + error_file: /tmp/torchelastic_uzuzl5wj/14018457_7qh_8ojw/attempt_0/2/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[3]: + time : 2025-01-02_00:41:14 + host : ip-26-0-163-43.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 97261) + error_file: /tmp/torchelastic_uzuzl5wj/14018457_7qh_8ojw/attempt_0/3/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[4]: + time : 2025-01-02_00:41:14 + host : ip-26-0-163-43.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 97262) + error_file: /tmp/torchelastic_uzuzl5wj/14018457_7qh_8ojw/attempt_0/4/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[5]: + time : 2025-01-02_00:41:14 + host : ip-26-0-163-43.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 97263) + error_file: /tmp/torchelastic_uzuzl5wj/14018457_7qh_8ojw/attempt_0/5/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[6]: + time : 2025-01-02_00:41:14 + host : ip-26-0-163-43.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 97264) + error_file: /tmp/torchelastic_uzuzl5wj/14018457_7qh_8ojw/attempt_0/6/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +[7]: + time : 2025-01-02_00:41:14 + host : ip-26-0-163-43.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 97265) + error_file: /tmp/torchelastic_uzuzl5wj/14018457_7qh_8ojw/attempt_0/7/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2025-01-02_00:41:14 + host : ip-26-0-163-43.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 97258) + error_file: /tmp/torchelastic_uzuzl5wj/14018457_7qh_8ojw/attempt_0/0/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 463, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 500, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 245, in train_batch_iter + model.p2p.clear_history() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__ + raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'") + AttributeError: 'DistributedDataParallel' object has no attribute 'p2p' + +============================================================ +srun: error: ip-26-0-163-43: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=14018457.0 +[2025-01-02 00:41:38,054] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 00:41:38,054] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125696 closing signal SIGTERM +[2025-01-02 00:41:38,054] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125697 closing signal SIGTERM +[2025-01-02 00:41:38,054] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125698 closing signal SIGTERM +[2025-01-02 00:41:38,054] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125699 closing signal SIGTERM +[2025-01-02 00:41:38,054] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125701 closing signal SIGTERM +[2025-01-02 00:41:38,054] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125702 closing signal SIGTERM +[2025-01-02 00:41:38,052] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 00:41:38,052] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46376 closing signal SIGTERM +[2025-01-02 00:41:38,052] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46377 closing signal SIGTERM +[2025-01-02 00:41:38,052] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46378 closing signal SIGTERM +[2025-01-02 00:41:38,052] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46379 closing signal SIGTERM +[2025-01-02 00:41:38,052] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46380 closing signal SIGTERM +[2025-01-02 00:41:38,052] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46381 closing signal SIGTERM +[2025-01-02 00:41:38,052] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 46382 closing signal SIGTERM +srun: error: ip-26-0-171-56: task 3: Terminated +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 125622 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 46304 got signal: 15 +srun: error: ip-26-0-171-21: task 2: Exited with exit code 1 +srun: error: ip-26-0-170-160: task 1: Exited with exit code 1 +srun: Force Terminated StepId=14018457.0 diff --git a/logs/14019228-bench_469G_dp64_tp8_pp1_acc2_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/14019228-bench_469G_dp64_tp8_pp1_acc2_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..9ac59170259d6c5fabf3811cbef7200097c4f13a --- /dev/null +++ b/logs/14019228-bench_469G_dp64_tp8_pp1_acc2_mbs2_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,17868 @@ ++ '[' -z 14019228 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-160-[100,103,192],ip-26-0-163-236,ip-26-0-164-[0,18,45,75,187,207,236],ip-26-0-165-[24,38,59,131,164,202,213],ip-26-0-166-[15,36,125,214,244],ip-26-0-167-[9,51,175,177,217,245],ip-26-0-168-[30,34,52,95,120,238],ip-26-0-169-[86,132,139,207,239,247],ip-26-0-170-[31,132,143,160],ip-26-0-171-[21,56,62,88,102,168,230,249],ip-26-0-172-[57,73,116,142,147,252],ip-26-0-173-[7,121,202,246],ip-26-0-174-36' ++ export 'NODELIST=ip-26-0-160-100 +ip-26-0-160-103 +ip-26-0-160-192 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36' ++ NODELIST='ip-26-0-160-100 +ip-26-0-160-103 +ip-26-0-160-192 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-160-[100,103,192],ip-26-0-163-236,ip-26-0-164-[0,18,45,75,187,207,236],ip-26-0-165-[24,38,59,131,164,202,213],ip-26-0-166-[15,36,125,214,244],ip-26-0-167-[9,51,175,177,217,245],ip-26-0-168-[30,34,52,95,120,238],ip-26-0-169-[86,132,139,207,239,247],ip-26-0-170-[31,132,143,160],ip-26-0-171-[21,56,62,88,102,168,230,249],ip-26-0-172-[57,73,116,142,147,252],ip-26-0-173-[7,121,202,246],ip-26-0-174-36' ++ export MASTER_NODE=ip-26-0-160-100 ++ MASTER_NODE=ip-26-0-160-100 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=64 ++ NNODES=64 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=512 ++ WORLD_SIZE=512 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-160-100' +Master node: ip-26-0-160-100 ++ echo 'All nodes: ip-26-0-160-100 +ip-26-0-160-103 +ip-26-0-160-192 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36' +All nodes: ip-26-0-160-100 +ip-26-0-160-103 +ip-26-0-160-192 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-207 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-102 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-36 ++ echo 'World size: 512' +World size: 512 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=64 --nproc_per_node=8 --rdzv_id=14019228 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-100:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_469G_dp64_tp8_pp1_acc2_mbs2_seq4096_zero1_tpmodeRED_vocab131k.yaml +[2025-01-02 20:16:10,275] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,277] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,277] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,277] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,277] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,277] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,275] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,275] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,275] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,275] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,276] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,278] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,278] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,278] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,278] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,278] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,279] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,279] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,279] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,279] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,279] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,279] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,280] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,282] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,282] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,282] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,282] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,282] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,281] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,284] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,284] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,284] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,284] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,284] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,286] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,282] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,283] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,287] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,285] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,288] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,289] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,289] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,289] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,289] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,290] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,290] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,290] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,290] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,290] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,292] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,293] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,293] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,293] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,293] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,293] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,291] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,293] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,293] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,293] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,293] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,295] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,295] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,295] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,295] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,295] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,295] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,296] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,296] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,296] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,296] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,300] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,300] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,300] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,300] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,300] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,301] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,302] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,302] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,302] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,302] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,303] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,304] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,304] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,304] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,304] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,304] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,304] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,304] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,304] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,304] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,308] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,323] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,323] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,323] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,323] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,323] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,327] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,327] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,327] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,327] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,327] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,338] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,338] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,338] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,338] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,338] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,308] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,308] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,308] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,308] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,353] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,353] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,353] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,353] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,353] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,354] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,355] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,355] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,355] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,355] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,355] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,355] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,355] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,355] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,355] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,370] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,370] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,370] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,370] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,370] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,374] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,374] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,374] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,374] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,374] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,342] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,507] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-02 20:16:10,507] torch.distributed.run: [WARNING] +[2025-01-02 20:16:10,507] torch.distributed.run: [WARNING] ***************************************** +[2025-01-02 20:16:10,507] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-02 20:16:10,507] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Config: +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Config(general=GeneralArgs(project='debug', +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: run='469G_dp64_tp8_pp1_acc2_mbs2_seq4096_zero1_tpmodeRED_vocab131k', +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: seed=42, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: step=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: consumed_train_samples=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: ignore_sanity_checks=True), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: parallelism=ParallelismArgs(dp=64, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pp=1, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp=8, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pp_engine=, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp_mode=, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp_linear_async_communication=True, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: recompute_layer=False, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp_recompute_allgather=True, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: expert_parallel_size=1), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: eos_token_id=0, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_act='silu', +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_size=16384, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: initializer_range=0.02, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: intermediate_size=53248, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: is_llama_config=True, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: max_position_embeddings=4096, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_attention_heads=128, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_hidden_layers=126, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_key_value_heads=128, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pad_token_id=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pretraining_tp=1, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rms_norm_eps=1e-05, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_scaling=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_theta=10000.0, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_interleaved=False, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tie_word_embeddings=False, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: use_cache=True, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: vocab_size=131072), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: init_method=RandomInit(std=0.02), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: dtype=torch.bfloat16, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: make_vocab_size_divisible_by=1, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: ddp_bucket_cap_mb=25), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokenizer_revision=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokenizer_max_length=None), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: checkpoint_interval=10000, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: save_initial_state=False, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: save_final_state=False, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: resume_checkpoint_path=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: checkpoints_path_is_shared_file_system=False), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: logging=LoggingArgs(log_level='info', +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: log_level_replica='info', +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: iteration_step_info_interval=1), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokens=TokensArgs(sequence_length=4096, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: train_steps=100, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: micro_batch_size=2, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: batch_accumulation_per_replica=2, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: val_check_interval=100, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: limit_val_batches=0, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: limit_test_batches=0), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: adam_beta1=0.9, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: adam_beta2=0.95, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: torch_adam_is_fused=True, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: name='adamW'), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: zero_stage=1, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: weight_decay=0.01, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: clip_grad=1.0, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: accumulate_grad_in_fp32=True, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_warmup_steps=2, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_warmup_style='linear', +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_decay_style='cosine', +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_decay_steps=13, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_decay_starting_step=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: min_decay_lr=1e-05)), +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: start_training_step=1, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: data=DataArgs(dataset=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: seed=42, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_loading_workers=1))], +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: profiler=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lighteval=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: s3_upload=None) +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Model Config: +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: LlamaConfig(bos_token_id=0, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: eos_token_id=0, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_act='silu', +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_size=16384, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: initializer_range=0.02, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: intermediate_size=53248, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: is_llama_config=True, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: max_position_embeddings=4096, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_attention_heads=128, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_hidden_layers=126, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_key_value_heads=128, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pad_token_id=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pretraining_tp=1, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rms_norm_eps=1e-05, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_scaling=None, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_theta=10000.0, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_interleaved=False, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tie_word_embeddings=False, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: use_cache=True, +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: vocab_size=131072) +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Building model.. +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Initialize RoPE Theta = 10000.0 +01/02/2025 20:17:48 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Setting PP block ranks... +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 71.22 GiB is allocated by PyTorch, and 112.72 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.qkv_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 35.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.qkv_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.attn = CausalSelfAttention( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + model = self._init_model_instance() + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = build_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.qkv_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 187.94 MiB is free. Including non-PyTorch memory, this process has 79.13 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__(model = self._init_model_instance() + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.attn = CausalSelfAttention( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + model = self._init_model_instance() + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + model = build_model( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.qkv_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + model = self._init_model_instance() + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + model = self._init_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + model = build_model( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + self.o_proj = TensorParallelRowLinear( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + model = build_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.qkv_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + model = build_model( + super().__init__( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.attn = CausalSelfAttention( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.qkv_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + self.model = self.init_model() # Defines self.model + super().__init__( File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.attn = CausalSelfAttention( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.o_proj = TensorParallelRowLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + model = self._init_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + model = build_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.model = self.init_model() # Defines self.model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + super().__init__( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + model = self._init_model_instance() + model = build_model( + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.qkv_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + model = build_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + return fn(*args, **kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 71.22 GiB is allocated by PyTorch, and 112.72 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.qkv_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + model = self._init_model_instance() + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + self.qkv_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + super().__init__( File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + + model = build_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs)self.pp_block = self.module_builder(**self.module_kwargs) + + model = self._init_model_instance() + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.qkv_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + model = build_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 35.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + model = self._init_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + model = build_model( + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.qkv_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.gate_up_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 187.94 MiB is free. Including non-PyTorch memory, this process has 79.13 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + model = self._init_model( + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + return fn(*args, **kwargs) + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + self.model = self.init_model() # Defines self.model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.model = self.init_model() # Defines self.model + model = build_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.o_proj = TensorParallelRowLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + model = self._init_model_instance() + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.model = self.init_model() # Defines self.model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + super().__init__( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + model = self._init_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + model = build_model( + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 227.94 MiB is free. Including non-PyTorch memory, this process has 79.10 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + self.qkv_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + self.o_proj = TensorParallelRowLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + trainer = DistributedTrainer(config_file) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + return fn(*args, **kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + model = build_model( + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model_instance() + self.gate_up_proj = TensorParallelColumnLinear( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + self.model = self.init_model() # Defines self.model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + model = self._init_model( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model_instance() + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.attn = CausalSelfAttention( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + self.qkv_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file)self.gate_up_proj = TensorParallelColumnLinear( + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + model = build_model( + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.attn = CausalSelfAttention( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.qkv_proj = TensorParallelColumnLinear( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.97 GiB is allocated by PyTorch, and 112.75 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + model = self._init_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + model = build_model( + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.o_proj = TensorParallelRowLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( +Traceback (most recent call last): + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model_instance() + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + super().__init__( + return fn(*args, **kwargs) + return fn(*args, **kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + trainer = DistributedTrainer(config_file) + super().__init__( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + self.model = self.init_model() # Defines self.model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + super().__init__( + super().__init__( + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + model = self._init_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.model = self.init_model() # Defines self.model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + model = build_model( + model = self._init_model_instance() + self.gate_up_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model_instance() + model = self._init_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + model = self._init_model( + return fn(*args, **kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( +Traceback (most recent call last): + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + model = self._init_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + return fn(*args, **kwargs) + return fn(*args, **kwargs) + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 115.94 MiB is free. Including non-PyTorch memory, this process has 79.21 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 75.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + model = build_model( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + self.attn = CausalSelfAttention( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.o_proj = TensorParallelRowLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + model = self._init_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + self.model = self.init_model() # Defines self.model + model = self._init_model_instance() + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + super().__init__( + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + model = self._init_model_instance() + model = self._init_model( +Traceback (most recent call last): + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + model = build_model( + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + model = self._init_model( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + super().__init__( + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + model = build_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + self.model = self.init_model() # Defines self.model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + model = build_model( + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + model = build_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + return fn(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 275.94 MiB is free. Including non-PyTorch memory, this process has 79.05 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + model = build_model( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.o_proj = TensorParallelRowLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + return fn(*args, **kwargs) + model = self._init_model_instance() + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + super().__init__( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + self.attn = CausalSelfAttention( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.o_proj = TensorParallelRowLinear( + return fn(*args, **kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + model = self._init_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + super().__init__( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + return fn(*args, **kwargs) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model_instance() + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file)self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.attn = CausalSelfAttention( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.model = self.init_model() # Defines self.model +super().__init__( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + model = build_model( + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + model = self._init_model( + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + model = self._init_model_instance() + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + model = self._init_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + self.model = self.init_model() # Defines self.model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + model = build_model( + model = self._init_model_instance() + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() +trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model( + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model_instance() + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) + super().__init__( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + model = self._init_model( + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + model = build_model( + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 195.94 MiB is free. Including non-PyTorch memory, this process has 79.13 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + super().__init__( + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 195.94 MiB is free. Including non-PyTorch memory, this process has 79.13 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + model = build_model( + model = build_model( + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + return fn(*args, **kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.attn = CausalSelfAttention( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 315.94 MiB is free. Including non-PyTorch memory, this process has 79.01 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + model = build_model( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + model = build_model( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + model = build_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + return fn(*args, **kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + return fn(*args, **kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = build_model( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = build_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.attn = CausalSelfAttention( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + self.model = self.init_model() # Defines self.model + super().__init__( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + super().__init__( + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + return fn(*args, **kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + self.o_proj = TensorParallelRowLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ +model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + super().__init__( + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model_instance() + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + model = build_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = build_model( + model = build_model( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + model = self._init_model_instance() + model = build_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.attn = CausalSelfAttention( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + self.model = self.init_model() # Defines self.model + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.o_proj = TensorParallelRowLinear( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + self.o_proj = TensorParallelRowLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model( + model = self._init_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + trainer = DistributedTrainer(config_file) + super().__init__( + model = build_model( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + self.attn = CausalSelfAttention( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.attn = CausalSelfAttention( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 115.94 MiB is free. Including non-PyTorch memory, this process has 79.21 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + super().__init__( + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + model = self._init_model_instance() + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + return fn(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 355.94 MiB is free. Including non-PyTorch memory, this process has 78.97 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + model = self._init_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 128.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file)self.pp_block = self.module_builder(**self.module_kwargs) + + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 128.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 96.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cudaTraceback (most recent call last): +.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( +Traceback (most recent call last): + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.gate_up_proj = TensorParallelColumnLinear( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 83.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) + super().__init__( + self.model = self.init_model() # Defines self.model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + self.model = self.init_model() # Defines self.model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 75.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + trainer = DistributedTrainer(config_file) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + model = self._init_model_instance() + self.attn = CausalSelfAttention( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + model = build_model( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + Traceback (most recent call last): + return fn(*args, **kwargs) +model = build_model( + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 75.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 75.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + super().__init__( + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + return fn(*args, **kwargs) + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + super().__init__( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 107.94 MiB is free. Including non-PyTorch memory, this process has 79.21 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) +torch.cuda. File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + model = self._init_model_instance() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + return fn(*args, **kwargs) + model = self._init_model_instance() +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 187.94 MiB is free. Including non-PyTorch memory, this process has 79.13 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + return fn(*args, **kwargs) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 259.94 MiB is free. Including non-PyTorch memory, this process has 79.06 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.model = self.init_model() # Defines self.model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 243.94 MiB is free. Including non-PyTorch memory, this process has 79.08 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 128.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + self.gate_up_proj = TensorParallelColumnLinear( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 275.94 MiB is free. Including non-PyTorch memory, this process has 79.05 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 75.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 171.94 MiB is free. Including non-PyTorch memory, this process has 79.15 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + trainer = DistributedTrainer(config_file) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 128.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.attn = CausalSelfAttention( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 384, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.qkv_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +Traceback (most recent call last): + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.11 GiB is allocated by PyTorch, and 128.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.pp_block = self.module_builder(**self.module_kwargs) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + self.pp_block = self.module_builder(**self.module_kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 128.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 96.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 107.94 MiB is free. Including non-PyTorch memory, this process has 79.21 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 128.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 131.94 MiB is free. Including non-PyTorch memory, this process has 79.19 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 123.94 MiB is free. Including non-PyTorch memory, this process has 79.20 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 128.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + return fn(*args, **kwargs) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 379.94 MiB is free. Including non-PyTorch memory, this process has 78.95 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + super().__init__( + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 187.94 MiB is free. Including non-PyTorch memory, this process has 79.13 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.gate_up_proj = TensorParallelColumnLinear( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + super().__init__( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 179.94 MiB is free. Including non-PyTorch memory, this process has 79.14 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + return fn(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 299.94 MiB is free. Including non-PyTorch memory, this process has 79.03 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 35.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 163.94 MiB is free. Including non-PyTorch memory, this process has 79.16 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = self._init_model( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 3.94 MiB is free. Including non-PyTorch memory, this process has 79.31 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + model = self._init_model( + super().__init__( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + model = self._init_model( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 705, in __init__ + self.attn = CausalSelfAttention( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 414, in __init__ + self.o_proj = TensorParallelRowLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 127, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 64.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.30 GiB is allocated by PyTorch, and 112.82 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 155.94 MiB is free. Including non-PyTorch memory, this process has 79.17 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.gate_up_proj = TensorParallelColumnLinear(self.model = self.init_model() # Defines self.model + + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) + model = build_model( +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 211.94 MiB is free. Including non-PyTorch memory, this process has 79.11 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) +trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 99.94 MiB is free. Including non-PyTorch memory, this process has 79.22 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + return fn(*args, **kwargs)Traceback (most recent call last): + + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 75.94 MiB is free. Including non-PyTorch memory, this process has 79.24 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 96.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 0 has a total capacty of 79.33 GiB of which 91.94 MiB is free. Including non-PyTorch memory, this process has 79.23 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 4 has a total capacty of 79.33 GiB of which 43.94 MiB is free. Including non-PyTorch memory, this process has 79.28 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 6 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 5 has a total capacty of 79.33 GiB of which 139.94 MiB is free. Including non-PyTorch memory, this process has 79.18 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 7 has a total capacty of 79.33 GiB of which 339.94 MiB is free. Including non-PyTorch memory, this process has 78.99 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 2 has a total capacty of 79.33 GiB of which 59.94 MiB is free. Including non-PyTorch memory, this process has 79.26 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 249, in + trainer = DistributedTrainer(config_file) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 171, in __init__ + self.model = self.init_model() # Defines self.model + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 720, in init_model + model = self._init_model_instance() + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 730, in _init_model_instance + model = self._init_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 804, in _init_model + model = build_model( + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 232, in build_model + block.build_and_set_rank(target_pp_ranks[target_pp_rank_idx], block_rank_counter) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/block.py", line 62, in build_and_set_rank + self.pp_block = self.module_builder(**self.module_kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 713, in __init__ + self.mlp = MLP(config=config, parallel_config=parallel_config, tp_pg=tp_pg) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/llama.py", line 223, in __init__ + self.gate_up_proj = TensorParallelColumnLinear( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/nn.py", line 65, in __init__ + super().__init__( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 96, in __init__ + self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs)) + File "/fsx/nouamane/projects/nanotron/src/nanotron/models/base.py", line 330, in wrapper + return fn(*args, **kwargs) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 416.00 MiB. GPU 3 has a total capacty of 79.33 GiB of which 19.94 MiB is free. Including non-PyTorch memory, this process has 79.30 GiB memory in use. Of the allocated memory 70.36 GiB is allocated by PyTorch, and 112.79 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[2025-01-02 20:18:07,778] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40517 closing signal SIGTERM +[2025-01-02 20:18:07,778] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40519 closing signal SIGTERM +[2025-01-02 20:18:07,778] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40520 closing signal SIGTERM +[2025-01-02 20:18:07,778] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40521 closing signal SIGTERM +[2025-01-02 20:18:07,778] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40522 closing signal SIGTERM +[2025-01-02 20:18:07,778] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40523 closing signal SIGTERM +[2025-01-02 20:18:07,782] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343150 closing signal SIGTERM +[2025-01-02 20:18:07,782] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343151 closing signal SIGTERM +[2025-01-02 20:18:07,783] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343152 closing signal SIGTERM +[2025-01-02 20:18:07,783] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343153 closing signal SIGTERM +[2025-01-02 20:18:07,783] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343155 closing signal SIGTERM +[2025-01-02 20:18:07,783] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343156 closing signal SIGTERM +[2025-01-02 20:18:07,783] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611457 closing signal SIGTERM +[2025-01-02 20:18:07,783] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611458 closing signal SIGTERM +[2025-01-02 20:18:07,783] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611459 closing signal SIGTERM +[2025-01-02 20:18:07,783] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611461 closing signal SIGTERM +[2025-01-02 20:18:07,783] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611462 closing signal SIGTERM +[2025-01-02 20:18:07,783] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611463 closing signal SIGTERM +[2025-01-02 20:18:07,785] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249223 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154695 closing signal SIGTERM +[2025-01-02 20:18:07,785] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249224 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154696 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154697 closing signal SIGTERM +[2025-01-02 20:18:07,785] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249227 closing signal SIGTERM +[2025-01-02 20:18:07,785] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249228 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154698 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154699 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154700 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162658 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162659 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162660 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162661 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162663 closing signal SIGTERM +[2025-01-02 20:18:07,786] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162664 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173073 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173074 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173075 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173076 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173077 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173078 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 146214 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 146215 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 146216 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 146218 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 146220 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50899 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50900 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50901 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50902 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50903 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50904 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50905 closing signal SIGTERM +[2025-01-02 20:18:07,788] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400722 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400723 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400725 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400726 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400727 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331371 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400728 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588935 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331372 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331373 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331374 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588936 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288081 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331375 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588937 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32749 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331376 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588938 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288082 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588939 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288083 closing signal SIGTERM +[2025-01-02 20:18:07,789] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588940 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288084 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32750 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288085 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288086 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32751 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32753 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32755 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104917 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423504 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104918 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104919 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423505 closing signal SIGTERM +[2025-01-02 20:18:07,791] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423506 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104920 closing signal SIGTERM +[2025-01-02 20:18:07,791] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423508 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104921 closing signal SIGTERM +[2025-01-02 20:18:07,791] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423509 closing signal SIGTERM +[2025-01-02 20:18:07,791] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423510 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104922 closing signal SIGTERM +[2025-01-02 20:18:07,790] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104923 closing signal SIGTERM +[2025-01-02 20:18:07,791] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 607758 closing signal SIGTERM +[2025-01-02 20:18:07,791] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 607759 closing signal SIGTERM +[2025-01-02 20:18:07,791] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 607760 closing signal SIGTERM +[2025-01-02 20:18:07,791] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 607761 closing signal SIGTERM +[2025-01-02 20:18:07,791] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 607762 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 101943 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 101944 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 101945 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 101946 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 101948 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472025 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472026 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61795 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472027 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472028 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61796 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472029 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61797 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472031 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61798 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61799 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61801 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166397 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137759 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166398 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166399 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166401 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137760 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166402 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792168 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166403 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137761 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137762 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137763 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792169 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792171 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792172 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792173 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137764 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792174 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137765 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338282 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338283 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338284 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338285 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338287 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82291 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338288 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 229393 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82292 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82293 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 229395 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 229396 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 292659) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37725 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82294 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 229398 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 229399 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37726 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82296 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102777 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37727 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82297 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37729 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162036 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102778 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37730 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102779 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162037 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162038 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37731 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162039 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162041 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102780 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102781 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102782 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494804 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494805 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494806 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494807 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494808 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604628 closing signal SIGTERM +[2025-01-02 20:18:07,794] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494809 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604629 closing signal SIGTERM +[2025-01-02 20:18:07,792] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105153 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604630 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604631 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604632 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105154 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105155 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105156 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604634 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105157 closing signal SIGTERM +[2025-01-02 20:18:07,793] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105158 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1057008 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104717 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1057009 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104718 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 294030 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104719 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 543677 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1057010 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1057012 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104720 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1057014 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 543678 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 294031 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104722 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 543679 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 543680 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 294032 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 543682 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 294033 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 294035 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 87497 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212633 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212634 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212638 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212639 closing signal SIGTERM +[2025-01-02 20:18:07,795] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514572 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514573 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514574 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514575 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514576 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514578 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135168 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135169 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135170 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135171 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135172 closing signal SIGTERM +[2025-01-02 20:18:07,796] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184460 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135174 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184461 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184462 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184463 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184464 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184465 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84956 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147694 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84957 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84958 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147695 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84960 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84961 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147697 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147698 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84962 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147699 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147700 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47195 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47196 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47197 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47198 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47199 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220906 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47201 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220908 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220909 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220910 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220911 closing signal SIGTERM +[2025-01-02 20:18:07,797] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220912 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786605 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551126 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786606 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551127 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570720 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551128 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786607 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551130 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551131 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570721 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570723 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786608 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551132 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570724 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570725 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570726 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786609 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786610 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90605 closing signal SIGTERM +[2025-01-02 20:18:07,798] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90607 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90608 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90609 closing signal SIGTERM +[2025-01-02 20:18:07,799] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41610 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41611 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41612 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41613 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41614 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41615 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252545 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252546 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252547 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252548 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252549 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252550 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166693 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166694 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166695 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166696 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166697 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67749 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166699 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67750 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67751 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67753 closing signal SIGTERM +[2025-01-02 20:18:07,800] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67754 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300207 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163491 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300208 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300209 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300210 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163492 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163493 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300211 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163494 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163495 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163496 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300212 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163497 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44029 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44030 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44031 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44032 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44033 closing signal SIGTERM +[2025-01-02 20:18:07,801] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44034 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30344 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30346 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30347 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30349 closing signal SIGTERM +[2025-01-02 20:18:07,802] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30350 closing signal SIGTERM +[2025-01-02 20:18:07,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 292702 closing signal SIGTERM +[2025-01-02 20:18:07,804] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 292703 closing signal SIGTERM +[2025-01-02 20:18:07,804] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 292704 closing signal SIGTERM +[2025-01-02 20:18:07,804] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 292706 closing signal SIGTERM +[2025-01-02 20:18:07,804] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 292707 closing signal SIGTERM +[2025-01-02 20:18:07,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3978138 closing signal SIGTERM +[2025-01-02 20:18:07,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3978139 closing signal SIGTERM +[2025-01-02 20:18:07,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3978140 closing signal SIGTERM +[2025-01-02 20:18:07,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3978141 closing signal SIGTERM +[2025-01-02 20:18:07,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3978142 closing signal SIGTERM +[2025-01-02 20:18:07,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125565 closing signal SIGTERM +[2025-01-02 20:18:07,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125566 closing signal SIGTERM +[2025-01-02 20:18:07,803] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125567 closing signal SIGTERM +[2025-01-02 20:18:07,804] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125568 closing signal SIGTERM +[2025-01-02 20:18:07,804] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125569 closing signal SIGTERM +[2025-01-02 20:18:07,804] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125571 closing signal SIGTERM +[2025-01-02 20:18:07,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330863 closing signal SIGTERM +[2025-01-02 20:18:07,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330864 closing signal SIGTERM +[2025-01-02 20:18:07,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330865 closing signal SIGTERM +[2025-01-02 20:18:07,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330866 closing signal SIGTERM +[2025-01-02 20:18:07,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330867 closing signal SIGTERM +[2025-01-02 20:18:07,806] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330868 closing signal SIGTERM +[2025-01-02 20:18:07,805] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163404 closing signal SIGTERM +[2025-01-02 20:18:07,805] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163405 closing signal SIGTERM +[2025-01-02 20:18:07,805] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163406 closing signal SIGTERM +[2025-01-02 20:18:07,805] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163408 closing signal SIGTERM +[2025-01-02 20:18:07,809] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71563 closing signal SIGTERM +[2025-01-02 20:18:07,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71564 closing signal SIGTERM +[2025-01-02 20:18:07,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71567 closing signal SIGTERM +[2025-01-02 20:18:07,810] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71568 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2025-01-02_20:18:07 + host : ip-26-0-174-36.ec2.internal + rank : 505 (local_rank: 1) + exitcode : 1 (pid: 292660) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2025-01-02_20:18:07 + host : ip-26-0-174-36.ec2.internal + rank : 506 (local_rank: 2) + exitcode : 1 (pid: 292661) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2025-01-02_20:18:07 + host : ip-26-0-174-36.ec2.internal + rank : 507 (local_rank: 3) + exitcode : 1 (pid: 292662) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2025-01-02_20:18:07 + host : ip-26-0-174-36.ec2.internal + rank : 508 (local_rank: 4) + exitcode : 1 (pid: 292663) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2025-01-02_20:18:07 + host : ip-26-0-174-36.ec2.internal + rank : 509 (local_rank: 5) + exitcode : 1 (pid: 292664) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2025-01-02_20:18:07 + host : ip-26-0-174-36.ec2.internal + rank : 510 (local_rank: 6) + exitcode : 1 (pid: 292665) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2025-01-02_20:18:07 + host : ip-26-0-174-36.ec2.internal + rank : 511 (local_rank: 7) + exitcode : 1 (pid: 292666) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2025-01-02_20:18:07 + host : ip-26-0-174-36.ec2.internal + rank : 504 (local_rank: 0) + exitcode : 1 (pid: 292659) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +[2025-01-02 20:18:08,111] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 87495) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2025-01-02_20:18:07 + host : ip-26-0-160-100.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 87496) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2025-01-02_20:18:07 + host : ip-26-0-160-100.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 87498) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2025-01-02_20:18:07 + host : ip-26-0-160-100.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 87499) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2025-01-02_20:18:07 + host : ip-26-0-160-100.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 87500) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2025-01-02_20:18:07 + host : ip-26-0-160-100.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 87501) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2025-01-02_20:18:07 + host : ip-26-0-160-100.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 87502) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2025-01-02_20:18:07 + host : ip-26-0-160-100.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 87495) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-174-36: task 63: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=14019228.0 +slurmstepd: error: *** STEP 14019228.0 ON ip-26-0-160-100 CANCELLED AT 2025-01-02T20:18:08 *** +[2025-01-02 20:18:08,228] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,228] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71563 closing signal SIGTERM +[2025-01-02 20:18:08,228] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71564 closing signal SIGTERM +[2025-01-02 20:18:08,228] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71567 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,228] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 71568 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343150 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343151 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343152 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343153 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343155 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 343156 closing signal SIGTERM +[2025-01-02 20:18:08,227] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,229] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288081 closing signal SIGTERM +[2025-01-02 20:18:08,227] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125565 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288082 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288083 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288084 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288085 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104917 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 288086 closing signal SIGTERM +[2025-01-02 20:18:08,227] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125566 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,227] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125567 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 294031 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 294032 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104919 closing signal SIGTERM +[2025-01-02 20:18:08,227] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125568 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 294033 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 294035 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104920 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338282 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82291 closing signal SIGTERM +[2025-01-02 20:18:08,228] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125569 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611457 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338283 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338284 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82292 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104921 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163491 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611458 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166397 closing signal SIGTERM +[2025-01-02 20:18:08,228] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 125571 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611459 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338285 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166398 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154695 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82293 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423504 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163492 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611461 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338287 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166399 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104922 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423505 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162036 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163493 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166401 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154696 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82294 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423506 closing signal SIGTERM +[2025-01-02 20:18:08,228] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163494 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611462 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 338288 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166402 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154697 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423508 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162037 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163404 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163495 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 611463 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154698 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82296 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423509 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162038 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147694 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166403 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82297 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162039 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147695 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154699 closing signal SIGTERM +[2025-01-02 20:18:08,229] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163408 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163496 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147697 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 154700 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 423510 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3978138 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 163497 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147698 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400722 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162041 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3978140 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147699 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400723 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3978142 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400725 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400726 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40517 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 147700 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792168 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551126 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40519 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104718 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400727 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792169 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 400728 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50899 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40520 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551127 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40521 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551128 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792171 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50900 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551130 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792172 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40522 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104720 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50901 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40523 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551131 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792173 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 551132 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2792174 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 104722 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50902 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300207 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50903 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300208 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 292702 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300209 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50904 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300210 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 292704 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249223 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 50905 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 292706 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300211 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249224 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330863 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 300212 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330864 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330865 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330866 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249228 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330867 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331371 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331372 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331373 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494804 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 330868 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184460 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331374 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494805 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47195 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331375 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494806 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173073 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184461 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494807 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494808 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84956 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47196 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173074 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 331376 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173075 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184462 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173076 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 494809 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84957 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47197 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84958 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184463 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90605 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 146214 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212633 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 543677 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184464 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 229393 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47198 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173077 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 146215 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 184465 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 229395 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90607 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 173078 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 146216 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212634 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84960 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 543678 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47199 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90608 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 146220 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212638 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41610 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 543679 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 229399 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135168 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 90609 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41611 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84961 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 543680 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30344 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 47201 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135169 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166693 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41612 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 84962 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30346 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166694 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41613 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 543682 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514572 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102777 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32749 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 101943 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135170 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166695 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514573 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102778 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 30350 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166696 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41614 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514574 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32750 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166697 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 41615 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514575 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102779 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 101944 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135171 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786605 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514576 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32751 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786606 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 101945 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166699 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786607 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102780 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135172 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786608 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 514578 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32753 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 135174 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786609 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 101946 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102781 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 102782 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 786610 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1057008 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32755 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37725 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 101948 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44029 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1057009 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37726 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220906 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 607758 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44030 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137759 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 607759 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1057010 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 607760 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37727 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220908 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570720 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 607761 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44031 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137760 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220909 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 607762 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137761 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220910 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570721 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137762 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220911 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570723 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37729 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44032 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137763 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1057012 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570724 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61795 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570725 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37730 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44033 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 220912 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 44034 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137764 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1057014 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472025 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 137765 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61796 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472026 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 570726 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 37731 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162658 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61797 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,234] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,232] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472027 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162659 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472028 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162660 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252545 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162661 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61798 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67749 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252546 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162663 closing signal SIGTERM +[2025-01-02 20:18:08,234] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604628 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472029 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252547 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67750 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 472031 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 162664 closing signal SIGTERM +[2025-01-02 20:18:08,234] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604629 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61799 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67751 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252548 closing signal SIGTERM +[2025-01-02 20:18:08,234] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604630 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 61801 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67753 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105153 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252549 closing signal SIGTERM +[2025-01-02 20:18:08,234] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604631 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67754 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105154 closing signal SIGTERM +[2025-01-02 20:18:08,230] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105155 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105156 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105157 closing signal SIGTERM +[2025-01-02 20:18:08,232] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252550 closing signal SIGTERM +[2025-01-02 20:18:08,234] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604632 closing signal SIGTERM +[2025-01-02 20:18:08,234] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 604634 closing signal SIGTERM +[2025-01-02 20:18:08,231] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 105158 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588935 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588936 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588937 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588938 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588939 closing signal SIGTERM +[2025-01-02 20:18:08,233] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 588940 closing signal SIGTERM +srun: error: ip-26-0-160-100: task 0: Terminated +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 212560 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 249148 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 292628 got signal: 15 + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 30269 got signal: 15 + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 71490 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 163332 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 104643 got signal: 15 +srun: error: ip-26-0-171-62: task 47: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 3977963 got signal: 15 +srun: error: ip-26-0-167-177: task 26: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 90531 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 146120 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 293958 got signal: 15 +srun: error: ip-26-0-171-88: task 48: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 607682 got signal: 15 +srun: error: ip-26-0-160-192: task 2: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1056935 got signal: 15 +srun: error: ip-26-0-166-214: task 21: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 543603 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 331297 got signal: 15 +srun: error: ip-26-0-166-244: task 22: Exited with exit code 1 +srun: error: ip-26-0-167-245: task 28: Exited with exit code 1 +srun: error: ip-26-0-167-217: task 27: Exited with exit code 1 +srun: error: ip-26-0-165-38: task 12: Exited with exit code 1 +srun: error: ip-26-0-169-239: task 39: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 229321 got signal: 15 +srun: error: ip-26-0-173-246: task 62: Exited with exit code 1 +srun: error: ip-26-0-164-45: task 6: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 101870 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 161963 got signal: 15 +srun: error: ip-26-0-170-31: task 41: Exited with exit code 1 +srun: error: ip-26-0-171-168: task 50: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 288005 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ +srun: error: ip-26-0-168-52: task 31: Exited with exit code 1 + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 82218 got signal: 15 + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 3252471 got signal: 15 +srun: error: ip-26-0-171-56: task 46: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 162585 got signal: 15 + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 105080 got signal: 15 +srun: error: ip-26-0-169-132: task 36: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 786533 got signal: 15 +srun: error: ip-26-0-165-59: task 13: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return f(*args, **kwargs) + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + run(args) + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + result = self._pcontext.wait(0) + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._poll() + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 84882 got signal: 15 + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 300134 got signal: 15 + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 41537 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run +srun: error: ip-26-0-164-207: task 9: Exited with exit code 1 + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 102703 got signal: 15 +srun: error: ip-26-0-173-121: task 60: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 184387 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + run(args) + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + result = f(*args, **kwargs) + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + result = self._pcontext.wait(0) + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 43956 got signal: 15 + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 220832 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper +srun: error: ip-26-0-165-24: task 11: Exited with exit code 1 + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 400649 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ +srun: error: ip-26-0-172-57: task 53: Exited with exit code 1 + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper +torch.distributed.elastic.multiprocessing.api.SignalException: Process 423421 got signal: 15 + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 166618 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + self._close(death_sig=death_sig, timeout=timeout) + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + handler.proc.wait(time_to_wait) + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + return self._wait(timeout=timeout) + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 588863 got signal: 15 + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run +torch.distributed.elastic.multiprocessing.api.SignalException: Process 61723 got signal: 15 + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait +srun: error: ip-26-0-169-207: task 38: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 67676 got signal: 15 + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2792095 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 570647 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main +srun: error: ip-26-0-171-230: task 51: Exited with exit code 1 + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 330791 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + sys.exit(main()) + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + sys.exit(main()) + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + run(args) + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + return self._poll() + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + self.close() # terminate all running procs + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 147621 got signal: 15 + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 47105 got signal: 15 + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 40445 got signal: 15 + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 32676 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 338208 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 551053 got signal: 15 + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 471952 got signal: 15 +srun: error: ip-26-0-168-120: task 33: Exited with exit code 1 +srun: error: ip-26-0-170-160: task 44: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 137686 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 166325 got signal: 15 +srun: error: ip-26-0-170-143: task 43: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait +srun: error: ip-26-0-168-34: task 30: Exited with exit code 1 + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 135096 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run +srun: error: ip-26-0-164-18: task 5: Exited with exit code 1 + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 37652 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = agent.run() +srun: error: ip-26-0-168-95: task 32: Exited with exit code 1 + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + time.sleep(delay) +srun: error: ip-26-0-171-249: task 52: Exited with exit code 1 + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 154621 got signal: 15 + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 494730 got signal: 15 +srun: error: ip-26-0-172-116: task 55: Exited with exit code 1 +srun: error: ip-26-0-170-132: task 42: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +srun: error: ip-26-0-171-21: task 45: Exited with exit code 1 + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers +srun: error: ip-26-0-166-36: task 19: Exited with exit code 1 + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 104833 got signal: 15 +srun: error: ip-26-0-171-102: task 49: Exited with exit code 1 +srun: error: ip-26-0-166-125: task 20: Exited with exit code 1 +srun: error: ip-26-0-167-51: task 24: Exited with exit code 1 +srun: error: ip-26-0-172-147: task 57: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 125492 got signal: 15 +srun: error: ip-26-0-173-7: task 59: Exited with exit code 1 +srun: error: ip-26-0-165-202: task 16: Exited with exit code 1 +srun: error: ip-26-0-167-9: task 23: Exited with exit code 1 +srun: error: ip-26-0-168-238: task 34: Exited with exit code 1 +srun: error: ip-26-0-160-103: task 1: Exited with exit code 1 +srun: error: ip-26-0-164-187: task 8: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 514499 got signal: 15 +srun: error: ip-26-0-172-252: task 58: Exited with exit code 1 +srun: error: ip-26-0-166-15: task 18: Exited with exit code 1 +srun: error: ip-26-0-169-86: task 35: Exited with exit code 1 +srun: error: ip-26-0-169-139: task 37: Exited with exit code 1 +srun: error: ip-26-0-165-164: task 15: Exited with exit code 1 +srun: error: ip-26-0-172-73: task 54: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 172999 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 604556 got signal: 15 +srun: error: ip-26-0-172-142: task 56: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 50827 got signal: 15 +srun: error: ip-26-0-164-0: task 4: Exited with exit code 1 +srun: error: ip-26-0-168-30: task 29: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 163419 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 343077 got signal: 15 +srun: error: ip-26-0-163-236: task 3: Exited with exit code 1 +srun: error: ip-26-0-169-247: task 40: Exited with exit code 1 +srun: error: ip-26-0-167-175: task 25: Exited with exit code 1 +srun: error: ip-26-0-173-202: task 61: Exited with exit code 1 +srun: error: ip-26-0-165-213: task 17: Exited with exit code 1 +srun: error: ip-26-0-164-236: task 10: Exited with exit code 1 +srun: error: ip-26-0-164-75: task 7: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 611383 got signal: 15 +srun: error: ip-26-0-165-131: task 14: Exited with exit code 1 +srun: Force Terminated StepId=14019228.0 diff --git a/logs/14099129-bench_8.86G_dp1_tp16_pp8_acc16_mbs16_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/14099129-bench_8.86G_dp1_tp16_pp8_acc16_mbs16_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..dd9a9cb577d6f92b83f92a05b92d608ac029cf5f --- /dev/null +++ b/logs/14099129-bench_8.86G_dp1_tp16_pp8_acc16_mbs16_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,2007 @@ ++ '[' -z 14099129 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n x ']' +++++ manpath ++++ [[ ! :/admin/home/nouamane/miniconda/envs/2-1-cu121/man:/admin/home/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/slurm/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/local/cuda-12.1/include:1:/usr/bin:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/opt/aws-ofi-nccl/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/lib:/usr/lib +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib/:1:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2:/opt/aws-ofi-nccl/lib:2:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/efa/lib:1:/usr/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/local/cuda-12.1/include:1:/usr/bin:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.fzf/bin:1:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/admin/home/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/admin/home/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/admin/home/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/admin/home/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /admin/home/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/admin/home/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/admin/home/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/admin/home/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/slurm/bin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-161-[138,142,153,178,221],ip-26-0-167-[51,217,245],ip-26-0-168-[30,34,52,95,120,238],ip-26-0-169-[86,132]' ++ export 'NODELIST=ip-26-0-161-138 +ip-26-0-161-142 +ip-26-0-161-153 +ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-167-51 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132' ++ NODELIST='ip-26-0-161-138 +ip-26-0-161-142 +ip-26-0-161-153 +ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-167-51 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-161-[138,142,153,178,221],ip-26-0-167-[51,217,245],ip-26-0-168-[30,34,52,95,120,238],ip-26-0-169-[86,132]' ++ export MASTER_NODE=ip-26-0-161-138 ++ MASTER_NODE=ip-26-0-161-138 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=16 ++ NNODES=16 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=128 ++ WORLD_SIZE=128 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-43 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 52-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 52-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 52-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 52-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-161-138' +Master node: ip-26-0-161-138 ++ echo 'All nodes: ip-26-0-161-138 +ip-26-0-161-142 +ip-26-0-161-153 +ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-167-51 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132' +All nodes: ip-26-0-161-138 +ip-26-0-161-142 +ip-26-0-161-153 +ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-167-51 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 ++ echo 'World size: 128' +World size: 128 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=16 --nproc_per_node=8 --rdzv_id=14099129 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-161-138:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_8.86G_dp1_tp16_pp8_acc16_mbs16_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2025-01-07 14:31:32,889] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,889] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,889] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,889] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,889] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,889] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,890] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,889] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,890] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,890] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,890] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,890] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,893] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,895] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,912] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,913] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,889] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,889] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,889] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,889] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,890] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,890] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,890] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,890] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,891] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,895] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,895] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,895] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,895] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,893] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,893] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,893] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,893] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,938] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,943] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,949] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:32,912] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,912] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,912] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,912] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,913] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,913] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,913] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,913] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,938] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,938] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,938] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,938] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,943] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,943] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,943] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,943] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,949] torch.distributed.run: [WARNING] +[2025-01-07 14:31:32,949] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:32,949] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:32,949] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:33,077] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:33,077] torch.distributed.run: [WARNING] +[2025-01-07 14:31:33,077] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:33,077] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:33,077] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:36,430] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 14:31:36,430] torch.distributed.run: [WARNING] +[2025-01-07 14:31:36,430] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 14:31:36,430] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 14:31:36,430] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Config: +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Config(general=GeneralArgs(project='debug', +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: run='8.86G_dp1_tp16_pp8_acc16_mbs16_seq4096_zero0_tpmodeRED_vocab131k', +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: seed=42, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: step=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: consumed_train_samples=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: ignore_sanity_checks=True), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: parallelism=ParallelismArgs(dp=1, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pp=8, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tp=16, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pp_engine=, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tp_mode=, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tp_linear_async_communication=True, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: recompute_layer=False, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tp_recompute_allgather=True, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: expert_parallel_size=1), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: eos_token_id=0, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: hidden_act='silu', +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: hidden_size=4096, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: initializer_range=0.02, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: intermediate_size=14336, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: is_llama_config=True, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: max_position_embeddings=4096, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_attention_heads=32, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_hidden_layers=32, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_key_value_heads=32, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pad_token_id=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pretraining_tp=1, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rms_norm_eps=1e-05, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_scaling=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_theta=10000.0, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_interleaved=False, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tie_word_embeddings=False, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: use_cache=True, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: vocab_size=131072), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: init_method=RandomInit(std=0.02), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: dtype=torch.bfloat16, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: make_vocab_size_divisible_by=1, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: ddp_bucket_cap_mb=25), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tokenizer_revision=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tokenizer_max_length=None), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: checkpoint_interval=10000, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: save_initial_state=False, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: save_final_state=False, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: resume_checkpoint_path=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: checkpoints_path_is_shared_file_system=False), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: logging=LoggingArgs(log_level='info', +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: log_level_replica='info', +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: iteration_step_info_interval=1), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tokens=TokensArgs(sequence_length=4096, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: train_steps=100, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: micro_batch_size=16, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: batch_accumulation_per_replica=16, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: val_check_interval=100, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: limit_val_batches=0, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: limit_test_batches=0), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: adam_beta1=0.9, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: adam_beta2=0.95, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: torch_adam_is_fused=True, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: name='adamW'), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: zero_stage=0, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: weight_decay=0.01, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: clip_grad=1.0, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: accumulate_grad_in_fp32=True, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lr_warmup_steps=2, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lr_warmup_style='linear', +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lr_decay_style='cosine', +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lr_decay_steps=13, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lr_decay_starting_step=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: min_decay_lr=1e-05)), +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: start_training_step=1, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: data=DataArgs(dataset=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: seed=42, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_loading_workers=1))], +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: profiler=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: lighteval=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: s3_upload=None) +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Model Config: +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: LlamaConfig(bos_token_id=0, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: eos_token_id=0, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: hidden_act='silu', +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: hidden_size=4096, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: initializer_range=0.02, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: intermediate_size=14336, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: is_llama_config=True, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: max_position_embeddings=4096, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_attention_heads=32, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_hidden_layers=32, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: num_key_value_heads=32, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pad_token_id=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: pretraining_tp=1, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rms_norm_eps=1e-05, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_scaling=None, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_theta=10000.0, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: rope_interleaved=False, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: tie_word_embeddings=False, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: use_cache=True, +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: vocab_size=131072) +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Building model.. +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Initialize RoPE Theta = 10000.0 +01/07/2025 14:32:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Setting PP block ranks... +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.token_position_embeddings | PP: 0/8 | Block rank: 0 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.0 | PP: 0/8 | Block rank: 1 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.1 | PP: 0/8 | Block rank: 2 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.2 | PP: 0/8 | Block rank: 3 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.3 | PP: 0/8 | Block rank: 4 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.4 | PP: 0/8 | Block rank: 5 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.5 | PP: 1/8 | Block rank: 0 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.6 | PP: 1/8 | Block rank: 1 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.7 | PP: 1/8 | Block rank: 2 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.8 | PP: 1/8 | Block rank: 3 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.9 | PP: 1/8 | Block rank: 4 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.10 | PP: 2/8 | Block rank: 0 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.11 | PP: 2/8 | Block rank: 1 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.12 | PP: 2/8 | Block rank: 2 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.13 | PP: 2/8 | Block rank: 3 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.14 | PP: 2/8 | Block rank: 4 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.15 | PP: 3/8 | Block rank: 0 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.16 | PP: 3/8 | Block rank: 1 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.17 | PP: 3/8 | Block rank: 2 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.18 | PP: 3/8 | Block rank: 3 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.19 | PP: 3/8 | Block rank: 4 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.20 | PP: 4/8 | Block rank: 0 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.21 | PP: 4/8 | Block rank: 1 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.22 | PP: 4/8 | Block rank: 2 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.23 | PP: 4/8 | Block rank: 3 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.24 | PP: 4/8 | Block rank: 4 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.25 | PP: 5/8 | Block rank: 0 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.26 | PP: 5/8 | Block rank: 1 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.27 | PP: 5/8 | Block rank: 2 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.28 | PP: 5/8 | Block rank: 3 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.29 | PP: 5/8 | Block rank: 4 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.30 | PP: 6/8 | Block rank: 0 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.decoder.31 | PP: 6/8 | Block rank: 1 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.final_layer_norm | PP: 7/8 | Block rank: 0 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.lm_head | PP: 7/8 | Block rank: 1 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: model.cast_to_fp32 | PP: 7/8 | Block rank: 2 +01/07/2025 14:32:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: module_name: loss | PP: 7/8 | Block rank: 3 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1NCCL version 2.18.6+cuda12.1 + +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Total number of parameters: 8.86G (16904.12MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=0|ip-26-0-168-120]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=0|ip-26-0-168-120]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: No checkpoint path provided. +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Parametrizing model parameters using StandardParametrizator +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=0|ip-26-0-168-52]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=0|ip-26-0-161-153]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=0|ip-26-0-161-221]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=1|ip-26-0-161-138]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=5|ip-26-0-161-138]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=0|ip-26-0-168-52]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=6|ip-26-0-161-138]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=3|ip-26-0-161-138]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=2|ip-26-0-161-138]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=0|ip-26-0-161-221]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=0|ip-26-0-161-153]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=1|ip-26-0-161-138]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=5|ip-26-0-161-138]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=7|ip-26-0-161-138]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=6|ip-26-0-161-138]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=4|ip-26-0-161-138]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=3|ip-26-0-161-138]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=2|ip-26-0-161-138]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=7|ip-26-0-161-138]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=4|ip-26-0-161-138]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=8|ip-26-0-161-142]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=8|ip-26-0-161-142]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=8|ip-26-0-168-238]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=0|ip-26-0-167-245]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=8|ip-26-0-168-238]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=0|ip-26-0-167-245]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=1|ip-26-0-161-221]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=1|ip-26-0-168-30]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=5|ip-26-0-169-132]: Local number of parameters: 33.6M (64.01MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=1|ip-26-0-168-120]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=5|ip-26-0-168-120]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=5|ip-26-0-161-221]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=5|ip-26-0-168-30]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=5|ip-26-0-161-153]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=5|ip-26-0-168-52]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=1|ip-26-0-168-52]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=1|ip-26-0-169-132]: Local number of parameters: 33.6M (64.01MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=2|ip-26-0-168-52]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=2|ip-26-0-168-120]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=1|ip-26-0-161-153]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=3|ip-26-0-168-30]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=2|ip-26-0-168-30]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=2|ip-26-0-161-221]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=3|ip-26-0-161-221]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=1|ip-26-0-168-120]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=1|ip-26-0-168-30]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=5|ip-26-0-169-132]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=1|ip-26-0-161-221]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=5|ip-26-0-168-120]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=2|ip-26-0-167-245]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=3|ip-26-0-168-52]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=3|ip-26-0-169-132]: Local number of parameters: 33.6M (64.01MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=5|ip-26-0-161-221]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=2|ip-26-0-161-153]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=3|ip-26-0-161-153]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=2|ip-26-0-169-132]: Local number of parameters: 33.6M (64.01MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=5|ip-26-0-168-30]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=5|ip-26-0-168-52]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=7|ip-26-0-161-221]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=5|ip-26-0-161-153]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=1|ip-26-0-168-52]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=3|ip-26-0-168-30]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=2|ip-26-0-168-52]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=1|ip-26-0-161-153]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=3|ip-26-0-161-221]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=3|ip-26-0-168-120]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=2|ip-26-0-168-120]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=2|ip-26-0-168-30]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=2|ip-26-0-161-221]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=1|ip-26-0-169-132]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=3|ip-26-0-168-52]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=7|ip-26-0-161-153]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=3|ip-26-0-169-132]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=2|ip-26-0-167-245]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=3|ip-26-0-161-153]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=7|ip-26-0-161-221]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=7|ip-26-0-167-245]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=2|ip-26-0-161-153]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=2|ip-26-0-169-132]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=7|ip-26-0-168-120]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=7|ip-26-0-161-153]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=3|ip-26-0-168-120]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=6|ip-26-0-161-221]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=7|ip-26-0-167-245]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=6|ip-26-0-161-153]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=6|ip-26-0-168-52]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=6|ip-26-0-168-120]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=4|ip-26-0-161-221]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=7|ip-26-0-168-52]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=4|ip-26-0-168-120]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=4|ip-26-0-168-52]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=4|ip-26-0-169-132]: Local number of parameters: 33.6M (64.01MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=7|ip-26-0-169-132]: Local number of parameters: 33.6M (64.01MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=6|ip-26-0-161-221]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=4|ip-26-0-168-30]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=7|ip-26-0-168-30]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=7|ip-26-0-168-120]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=6|ip-26-0-161-153]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=6|ip-26-0-168-52]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=6|ip-26-0-168-120]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=4|ip-26-0-161-153]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=7|ip-26-0-168-52]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=4|ip-26-0-161-221]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=7|ip-26-0-169-132]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=4|ip-26-0-168-52]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=4|ip-26-0-168-120]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=4|ip-26-0-168-30]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=4|ip-26-0-169-132]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=7|ip-26-0-168-30]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=0|ip-26-0-169-132]: Local number of parameters: 33.6M (64.01MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=4|ip-26-0-161-153]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 9494.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=0|ip-26-0-168-30]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=0|ip-26-0-169-132]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=11|ip-26-0-161-142]: Local number of parameters: 110M (209.08MiB) +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=10|ip-26-0-161-142]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=0|ip-26-0-168-30]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=9|ip-26-0-161-142]: Local number of parameters: 110M (209.08MiB) +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=8|ip-26-0-168-95]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=11|ip-26-0-161-142]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=1|ip-26-0-167-245]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=10|ip-26-0-161-142]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=5|ip-26-0-167-245]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=9|ip-26-0-161-142]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=8|ip-26-0-168-34]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=8|ip-26-0-168-95]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=3|ip-26-0-167-245]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=1|ip-26-0-167-245]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=5|ip-26-0-167-245]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=3|ip-26-0-167-245]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=8|ip-26-0-168-34]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=8|ip-26-0-169-86]: Local number of parameters: 33.6M (64.01MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=14|ip-26-0-161-142]: Local number of parameters: 110M (209.08MiB) +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=6|ip-26-0-167-245]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=6|ip-26-0-169-132]: Local number of parameters: 33.6M (64.01MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=6|ip-26-0-168-30]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=8|ip-26-0-169-86]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=4|ip-26-0-167-245]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=13|ip-26-0-161-142]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=14|ip-26-0-161-142]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=8|ip-26-0-167-217]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=6|ip-26-0-167-245]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=15|ip-26-0-161-142]: Local number of parameters: 110M (209.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=6|ip-26-0-169-132]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=6|ip-26-0-168-30]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=12|ip-26-0-161-142]: Local number of parameters: 110M (209.08MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1NCCL version 2.18.6+cuda12.1 + +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=4|ip-26-0-167-245]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=8|ip-26-0-167-217]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 11542.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=13|ip-26-0-161-142]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=12|ip-26-0-161-142]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=0|TP=15|ip-26-0-161-142]: [After model building] Memory usage: 211.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=10|ip-26-0-168-238]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=11|ip-26-0-168-238]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=8|ip-26-0-161-178]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=10|ip-26-0-168-238]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=9|ip-26-0-168-238]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=9|ip-26-0-167-217]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=11|ip-26-0-168-238]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=9|ip-26-0-167-51]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=10|ip-26-0-167-217]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=8|ip-26-0-161-178]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=9|ip-26-0-168-238]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=10|ip-26-0-167-51]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=9|ip-26-0-167-217]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 11542.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=10|ip-26-0-168-95]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=9|ip-26-0-167-51]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=11|ip-26-0-168-95]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=11|ip-26-0-167-51]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=11|ip-26-0-167-217]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=10|ip-26-0-167-51]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=10|ip-26-0-167-217]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=10|ip-26-0-168-95]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=9|ip-26-0-168-95]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=11|ip-26-0-168-95]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=11|ip-26-0-167-51]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=11|ip-26-0-167-217]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=9|ip-26-0-168-95]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=8|ip-26-0-167-51]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=14|ip-26-0-168-238]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=13|ip-26-0-168-95]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=13|ip-26-0-161-178]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=14|ip-26-0-167-51]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=13|ip-26-0-167-217]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=13|ip-26-0-168-238]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=14|ip-26-0-167-217]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=8|ip-26-0-167-51]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=14|ip-26-0-168-238]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=13|ip-26-0-168-95]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=13|ip-26-0-161-178]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=15|ip-26-0-168-238]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=14|ip-26-0-167-51]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=13|ip-26-0-167-217]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=13|ip-26-0-168-238]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=14|ip-26-0-167-217]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=12|ip-26-0-168-238]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=15|ip-26-0-168-238]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=14|ip-26-0-168-95]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=4|TP=12|ip-26-0-168-238]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=14|ip-26-0-168-95]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=9|ip-26-0-161-178]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=15|ip-26-0-168-95]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=12|ip-26-0-168-95]: Local number of parameters: 30.4M (58.03MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=12|ip-26-0-167-51]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=12|ip-26-0-167-217]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=9|ip-26-0-161-178]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=12|ip-26-0-168-95]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=12|ip-26-0-167-51]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=6|TP=15|ip-26-0-168-95]: [After model building] Memory usage: 58.04MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=10|ip-26-0-161-178]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=12|ip-26-0-167-217]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=10|ip-26-0-168-34]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=15|ip-26-0-167-217]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=10|ip-26-0-161-178]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=11|ip-26-0-168-34]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=10|ip-26-0-168-34]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=9|ip-26-0-168-34]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=2|TP=15|ip-26-0-167-217]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 11542.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=11|ip-26-0-168-34]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=11|ip-26-0-161-178]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=9|ip-26-0-168-34]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=15|ip-26-0-167-51]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=11|ip-26-0-161-178]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=13|ip-26-0-168-34]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=14|ip-26-0-161-178]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=15|ip-26-0-167-51]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=13|ip-26-0-168-34]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=14|ip-26-0-161-178]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=14|ip-26-0-168-34]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=15|ip-26-0-168-34]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=10|ip-26-0-169-86]: Local number of parameters: 33.6M (64.01MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=14|ip-26-0-168-34]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=11|ip-26-0-169-86]: Local number of parameters: 33.6M (64.01MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=12|ip-26-0-161-178]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=12|ip-26-0-168-34]: Local number of parameters: 76.1M (145.08MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=15|ip-26-0-168-34]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=10|ip-26-0-169-86]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=11|ip-26-0-169-86]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=9|ip-26-0-169-86]: Local number of parameters: 33.6M (64.01MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=12|ip-26-0-161-178]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=5|TP=12|ip-26-0-168-34]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=15|ip-26-0-161-178]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=1|TP=15|ip-26-0-161-178]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=14|ip-26-0-169-86]: Local number of parameters: 33.6M (64.01MiB) +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=9|ip-26-0-169-86]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=13|ip-26-0-169-86]: Local number of parameters: 33.6M (64.01MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=14|ip-26-0-169-86]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=13|ip-26-0-167-51]: Local number of parameters: 76.1M (145.08MiB) +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=13|ip-26-0-169-86]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=15|ip-26-0-169-86]: Local number of parameters: 33.6M (64.01MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=3|TP=13|ip-26-0-167-51]: [After model building] Memory usage: 146.09MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=15|ip-26-0-169-86]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=12|ip-26-0-169-86]: Local number of parameters: 33.6M (64.01MiB) +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:36 [INFO|DP=0|PP=7|TP=12|ip-26-0-169-86]: [After model building] Memory usage: 64.02MiB. Peak allocated: 5392.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +NCCL version 2.18.6+cuda12.1 +01/07/2025 14:32:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [Optimizer Building] Using LearningRateForSP as learning rate +01/07/2025 14:32:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +01/07/2025 14:32:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Using dummy data generator +01/07/2025 14:32:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [Training Plan] There are 1 training stages +01/07/2025 14:32:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [Stage Stable Training Stage] start from step 1 +01/07/2025 14:32:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: +01/07/2025 14:32:37 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: [Start training] datetime: 2025-01-07 14:32:37.825875 | mbs: 16 | grad_accum: 16 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +01/07/2025 14:32:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +01/07/2025 14:32:38 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 1047.40MiB. Peak allocated 5392.00MiB. Peak reserved: 10518.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +01/07/2025 14:33:09 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 2140.97MiB. Peak allocated 33495.74MiB. Peak reserved: 59622.00MiB +01/07/2025 14:33:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 2977.30MiB. Peak allocated 2977.30MiB. Peak reserved: 59622.00MiB +01/07/2025 14:33:19 [INFO|DP=0|PP=7|TP=0|ip-26-0-169-132]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 40.1K | tokens_per_sec: 26.2K | tokens_per_sec_per_gpu: 204 | global_batch_size: 256 | lm_loss: 12.6 | lr: 0.00015 | model_tflops_per_gpu: 11.5 | hardware_tflops_per_gpu: 11.5 | grad_norm: 1.45 | cuda_memory_allocated: 1.21G | cuda_max_memory_reserved: 17.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 78.1G | hd_free_memory_tb: 234G +01/07/2025 14:33:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 2977.30MiB. Peak allocated 34332.07MiB. Peak reserved: 59622.00MiB +01/07/2025 14:33:24 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 2977.30MiB. Peak allocated 2977.32MiB. Peak reserved: 59622.00MiB +01/07/2025 14:33:24 [INFO|DP=0|PP=7|TP=0|ip-26-0-169-132]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 4.61K | tokens_per_sec: 228K | tokens_per_sec_per_gpu: 1.78K | global_batch_size: 256 | lm_loss: 12.6 | lr: 0.0003 | model_tflops_per_gpu: 100 | hardware_tflops_per_gpu: 100 | grad_norm: 1.45 | cuda_memory_allocated: 1.21G | cuda_max_memory_reserved: 17.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 78.1G | hd_free_memory_tb: 234G +01/07/2025 14:33:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 2977.30MiB. Peak allocated 34332.07MiB. Peak reserved: 59622.00MiB +01/07/2025 14:33:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 2977.30MiB. Peak allocated 2977.32MiB. Peak reserved: 59622.00MiB +01/07/2025 14:33:28 [INFO|DP=0|PP=7|TP=0|ip-26-0-169-132]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 4.64K | tokens_per_sec: 226K | tokens_per_sec_per_gpu: 1.76K | global_batch_size: 256 | lm_loss: 12.6 | lr: 0.000296 | model_tflops_per_gpu: 99.5 | hardware_tflops_per_gpu: 99.5 | grad_norm: 1.42 | cuda_memory_allocated: 1.21G | cuda_max_memory_reserved: 17.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 78.1G | hd_free_memory_tb: 234G +01/07/2025 14:33:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Memory usage: 2977.30MiB. Peak allocated 34332.07MiB. Peak reserved: 59622.00MiB +01/07/2025 14:33:33 [INFO|DP=0|PP=7|TP=0|ip-26-0-169-132]: iteration: 4 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 4.69K | tokens_per_sec: 223K | tokens_per_sec_per_gpu: 1.74K | global_batch_size: 256 | lm_loss: 12.5 | lr: 0.000283 | model_tflops_per_gpu: 98.4 | hardware_tflops_per_gpu: 98.4 | grad_norm: 0.944 | cuda_memory_allocated: 1.21G | cuda_max_memory_reserved: 17.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 78.1G | hd_free_memory_tb: 234G +01/07/2025 14:33:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +01/07/2025 14:33:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: | -------- | ---------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | ----------------------------------- | -------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +01/07/2025 14:33:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: | 14099129 | 8.86G_dp1_tp16_pp8_acc16_mbs16_seq4096_zero0_tpmodeRED_vocab131k | 16 | 4096 | 16 | 16 | 256 | 98.35 | 98.35 | 1744.67 | 48.58 | 25.93 | 25.45 | 426.86 | 242.76 | 242.11 | 2.91 | 58.22 | 1 | 8 | 16 | OneForwardOneBackwardPipelineEngine | REDUCE_SCATTER | True | 4096 | silu | 32 | 32 | 32 | 4096 | 131072 | False | torch.bfloat16 | 0 | 25 | True | 8.86G | 110M | +01/07/2025 14:33:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final2.csv +01/07/2025 14:33:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Acquired lock for CSV file: benchmark/results/bench_final2.csv +01/07/2025 14:33:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Successfully wrote to CSV file: benchmark/results/bench_final2.csv. Releasing lock... +01/07/2025 14:33:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-138]: Throughput logging complete +double free or corruption (!prev) +terminate called after throwing an instance of 'c10::Error' + what(): CUDA driver error: unknown error +Exception raised from _hasPrimaryContext at ../aten/src/ATen/cuda/detail/CUDAHooks.cpp:67 (most recent call first): +frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fbc3c267617 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10.so) +frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, char const*) + 0x68 (0x7fbc3c222a56 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10.so) +frame #2: + 0x11a00bf (0x7fbc3d5830bf in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #3: c10::cuda::MaybeSetDevice(int) + 0xc (0x7fbc3c318bdc in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libc10_cuda.so) +frame #4: std::_Sp_counted_ptr_inplace >, std::allocator > >, (__gnu_cxx::_Lock_policy)2>::_M_dispose() + 0x98 (0x7fbc3d692ae8 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #5: std::_Sp_counted_base<(__gnu_cxx::_Lock_policy)2>::_M_release() + 0x48 (0x7fbc9b5e5048 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_python.so) +frame #6: c10d::ProcessGroupNCCL::WorkNCCL::~WorkNCCL() + 0x135 (0x7fbc3d65dc55 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #7: c10d::ProcessGroupNCCL::workCleanupLoop() + 0x3c5 (0x7fbc3d673585 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #8: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x78 (0x7fbc3d673718 in /fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/lib/libtorch_cuda.so) +frame #9: + 0xe6793 (0x7fbc9e1b4793 in /lib/x86_64-linux-gnu/libstdc++.so.6) +frame #10: + 0x8609 (0x7fbccb85e609 in /lib/x86_64-linux-gnu/libpthread.so.0) +frame #11: clone + 0x43 (0x7fbccb629353 in /lib/x86_64-linux-gnu/libc.so.6) + +[2025-01-07 14:33:57,484] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 733813 closing signal SIGTERM +[2025-01-07 14:33:57,484] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 733814 closing signal SIGTERM +[2025-01-07 14:33:57,484] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 733816 closing signal SIGTERM +[2025-01-07 14:33:57,484] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 733817 closing signal SIGTERM +[2025-01-07 14:33:57,484] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 733818 closing signal SIGTERM +[2025-01-07 14:33:57,484] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 733819 closing signal SIGTERM +[2025-01-07 14:33:57,484] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 733820 closing signal SIGTERM +[2025-01-07 14:34:00,416] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: -6) local_rank: 2 (pid: 733815) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +======================================================= +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------- +Failures: + +------------------------------------------------------- +Root Cause (first observed failure): +[0]: + time : 2025-01-07_14:33:57 + host : ip-26-0-169-86.ec2.internal + rank : 122 (local_rank: 2) + exitcode : -6 (pid: 733815) + error_file: + traceback : Signal 6 (SIGABRT) received by PID 733815 +======================================================= +srun: error: ip-26-0-169-86: task 14: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=14099129.0 +slurmstepd: error: *** STEP 14099129.0 ON ip-26-0-161-138 CANCELLED AT 2025-01-07T14:34:00 *** +[2025-01-07 14:34:00,721] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,721] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155217 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155218 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155219 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 175135 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 175136 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 175137 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155220 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155221 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 175138 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 175139 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 175140 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155222 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 175141 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155223 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 155224 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 175142 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 156781 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 156782 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 156783 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 156784 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 156785 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1030861 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 156786 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 113497 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1030862 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 156787 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 156788 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 113498 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 113499 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1486255 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1486256 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 792329 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 113500 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1486257 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1486258 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 113501 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 113502 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1486259 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,720] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1486260 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 792330 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 113503 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1486261 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1486262 closing signal SIGTERM +[2025-01-07 14:34:00,720] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177771 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 49212 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 792331 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 113504 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67866 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177772 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 49213 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 792332 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67867 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78054 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78055 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177773 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 49214 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177774 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 792333 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67868 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2159858 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78056 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177775 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78057 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177776 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 49215 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67869 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 792334 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2159859 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177777 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2159860 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78058 closing signal SIGTERM +[2025-01-07 14:34:00,721] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 177778 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 49216 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78059 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 49217 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67870 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 792335 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67871 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124579 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 792336 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2159861 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78060 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 49218 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124580 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67872 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31473 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124581 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2159862 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78061 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 49219 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124582 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2159863 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67873 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31474 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124583 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2159864 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2159865 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31475 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31476 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124584 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124585 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31477 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31478 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124586 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31479 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 31480 closing signal SIGTERM +[2025-01-07 14:34:00,722] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 14:34:00,723] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108082 closing signal SIGTERM +[2025-01-07 14:34:00,723] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108083 closing signal SIGTERM +[2025-01-07 14:34:00,723] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108084 closing signal SIGTERM +[2025-01-07 14:34:00,723] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108085 closing signal SIGTERM +[2025-01-07 14:34:00,723] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108086 closing signal SIGTERM +[2025-01-07 14:34:00,723] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108087 closing signal SIGTERM +[2025-01-07 14:34:00,723] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108088 closing signal SIGTERM +[2025-01-07 14:34:00,723] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 108089 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1030778 got signal: 15 +srun: error: ip-26-0-161-138: task 0: Exited with exit code 1 +[2025-01-07 14:34:01,845] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-51.ec2.internal_77980_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,014] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-178.ec2.internal_155143_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,039] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-221.ec2.internal_175061_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,067] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-132.ec2.internal_177697_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,079] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-95.ec2.internal_124505_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,169] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-34.ec2.internal_49138_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,209] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-142.ec2.internal_113422_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,221] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-52.ec2.internal_792255_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,245] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-30.ec2.internal_67791_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,256] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-153.ec2.internal_1486179_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,279] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-120.ec2.internal_108008_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,281] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-238.ec2.internal_31399_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,321] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-245.ec2.internal_156707_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:02,367] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-217.ec2.internal_2159782_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 177697 got signal: 15 +[2025-01-07 14:34:06,847] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-51.ec2.internal_77980_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:07,015] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-178.ec2.internal_155143_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:07,040] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-221.ec2.internal_175061_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-169-132: task 15: Exited with exit code 1 +[2025-01-07 14:34:07,081] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-95.ec2.internal_124505_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:07,171] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-34.ec2.internal_49138_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:07,211] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-142.ec2.internal_113422_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:07,223] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-52.ec2.internal_792255_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:07,246] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-30.ec2.internal_67791_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:07,257] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-153.ec2.internal_1486179_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:07,281] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-120.ec2.internal_108008_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:07,282] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-238.ec2.internal_31399_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:07,323] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-245.ec2.internal_156707_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +[2025-01-07 14:34:07,368] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-217.ec2.internal_2159782_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 113422 got signal: 15 +srun: error: ip-26-0-161-142: task 1: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 792255 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 155143 got signal: 15 +srun: error: ip-26-0-168-52: task 10: Exited with exit code 1 +srun: error: ip-26-0-161-178: task 3: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 49138 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 156707 got signal: 15 +srun: error: ip-26-0-168-34: task 9: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 67791 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 124505 got signal: 15 +srun: error: ip-26-0-167-245: task 7: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 175061 got signal: 15 +srun: error: ip-26-0-168-95: task 11: Exited with exit code 1 +srun: error: ip-26-0-168-30: task 8: Exited with exit code 1 +srun: error: ip-26-0-161-221: task 4: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 108008 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2159782 got signal: 15 +srun: error: ip-26-0-168-120: task 12: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 77980 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 31399 got signal: 15 +srun: error: ip-26-0-167-217: task 6: Exited with exit code 1 +srun: error: ip-26-0-167-51: task 5: Exited with exit code 1 +srun: error: ip-26-0-168-238: task 13: Exited with exit code 1 +[2025-01-07 14:34:12,259] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-161-153.ec2.internal_1486179_0' has failed to send a keep-alive heartbeat to the rendezvous '14099129' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1486179 got signal: 15 +srun: error: ip-26-0-161-153: task 2: Exited with exit code 1 diff --git a/logs/14099174-bench_1.34G_dp16_tp8_pp2_acc1_mbs16_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/14099174-bench_1.34G_dp16_tp8_pp2_acc1_mbs16_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..a5d1f57f3a2114a26057d3216655b803c23acd50 --- /dev/null +++ b/logs/14099174-bench_1.34G_dp16_tp8_pp2_acc1_mbs16_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:106c23414fc63a4424774ff4696365334b393f9bd078dd67f577f5adc46404f3 +size 13447226 diff --git a/logs/14099259-bench_3.57G_dp8_tp4_pp8_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out b/logs/14099259-bench_3.57G_dp8_tp4_pp8_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..39baba9b392a91d73b7e674d3a1fcdfedc38eb86 --- /dev/null +++ b/logs/14099259-bench_3.57G_dp8_tp4_pp8_acc32_mbs1_seq4096_zero1_tpmodeRED_vocab131k.out @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:57898d7e3cb88f05fd47d60d505ef0a660de8c67d41a4aa7aafeecacc31fdbc1 +size 17001887 diff --git a/logs/14113419-bench_3.57G_dp8_tp4_pp1_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/14113419-bench_3.57G_dp8_tp4_pp1_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..c716484f6ea6ea4d6ea925b1736266d54e947634 --- /dev/null +++ b/logs/14113419-bench_3.57G_dp8_tp4_pp1_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,886 @@ ++ echo 'Running script: /var/spool/slurmd/job14113419/slurm_script' +Running script: /var/spool/slurmd/job14113419/slurm_script ++ '[' -z 14113419 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-172-252,ip-26-0-173-[7,121,202]' ++ export 'NODELIST=ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202' ++ NODELIST='ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-172-252,ip-26-0-173-[7,121,202]' ++ export MASTER_NODE=ip-26-0-172-252 ++ MASTER_NODE=ip-26-0-172-252 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=4 ++ NNODES=4 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=32 ++ WORLD_SIZE=32 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-172-252' +Master node: ip-26-0-172-252 ++ echo 'All nodes: ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202' +All nodes: ip-26-0-172-252 +ip-26-0-173-7 +ip-26-0-173-121 +ip-26-0-173-202 ++ echo 'World size: 32' +World size: 32 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=4 --nproc_per_node=8 --rdzv_id=14113419 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-172-252:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_3.57G_dp8_tp4_pp1_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2025-01-07 16:50:35,862] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 16:50:35,867] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 16:50:35,870] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 16:50:35,875] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-07 16:50:35,862] torch.distributed.run: [WARNING] +[2025-01-07 16:50:35,862] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 16:50:35,862] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 16:50:35,862] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 16:50:35,870] torch.distributed.run: [WARNING] +[2025-01-07 16:50:35,870] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 16:50:35,870] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 16:50:35,870] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 16:50:35,867] torch.distributed.run: [WARNING] +[2025-01-07 16:50:35,867] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 16:50:35,867] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 16:50:35,867] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 16:50:35,875] torch.distributed.run: [WARNING] +[2025-01-07 16:50:35,875] torch.distributed.run: [WARNING] ***************************************** +[2025-01-07 16:50:35,875] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-07 16:50:35,875] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Config: +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Config(general=GeneralArgs(project='debug', +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: run='3.57G_dp8_tp4_pp1_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k', +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: seed=42, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: step=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: consumed_train_samples=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: ignore_sanity_checks=True), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: parallelism=ParallelismArgs(dp=8, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: pp=1, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: tp=4, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: pp_engine=, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: tp_mode=, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: tp_linear_async_communication=True, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: recompute_layer=False, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: tp_recompute_allgather=True, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: expert_parallel_size=1), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: eos_token_id=0, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: hidden_act='silu', +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: hidden_size=3072, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: initializer_range=0.02, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: intermediate_size=8192, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: is_llama_config=True, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: max_position_embeddings=4096, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: num_attention_heads=32, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: num_hidden_layers=28, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: num_key_value_heads=32, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: pad_token_id=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: pretraining_tp=1, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: rms_norm_eps=1e-05, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: rope_scaling=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: rope_theta=10000.0, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: rope_interleaved=False, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: tie_word_embeddings=True, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: use_cache=True, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: vocab_size=131072), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: init_method=RandomInit(std=0.02), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: dtype=torch.bfloat16, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: make_vocab_size_divisible_by=1, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: ddp_bucket_cap_mb=25), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: tokenizer_revision=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: tokenizer_max_length=None), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: checkpoint_interval=10000, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: save_initial_state=False, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: save_final_state=False, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: resume_checkpoint_path=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: checkpoints_path_is_shared_file_system=False), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: logging=LoggingArgs(log_level='info', +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: log_level_replica='info', +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: iteration_step_info_interval=1), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: tokens=TokensArgs(sequence_length=4096, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: train_steps=100, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: micro_batch_size=8, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: batch_accumulation_per_replica=4, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: val_check_interval=100, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: limit_val_batches=0, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: limit_test_batches=0), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: adam_beta1=0.9, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: adam_beta2=0.95, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: torch_adam_is_fused=True, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: name='adamW'), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: zero_stage=0, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: weight_decay=0.01, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: clip_grad=1.0, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: accumulate_grad_in_fp32=True, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: lr_warmup_steps=2, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: lr_warmup_style='linear', +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: lr_decay_style='cosine', +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: lr_decay_steps=13, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: lr_decay_starting_step=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: min_decay_lr=1e-05)), +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: start_training_step=1, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: data=DataArgs(dataset=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: seed=42, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: num_loading_workers=1))], +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: profiler=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: lighteval=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: s3_upload=None) +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Model Config: +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: LlamaConfig(bos_token_id=0, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: eos_token_id=0, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: hidden_act='silu', +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: hidden_size=3072, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: initializer_range=0.02, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: intermediate_size=8192, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: is_llama_config=True, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: max_position_embeddings=4096, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: num_attention_heads=32, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: num_hidden_layers=28, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: num_key_value_heads=32, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: pad_token_id=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: pretraining_tp=1, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: rms_norm_eps=1e-05, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: rope_scaling=None, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: rope_theta=10000.0, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: rope_interleaved=False, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: tie_word_embeddings=True, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: use_cache=True, +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: vocab_size=131072) +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Building model.. +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Initialize RoPE Theta = 10000.0 +01/07/2025 16:51:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Setting PP block ranks... +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.token_position_embeddings | PP: 0/1 | Block rank: 0 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.0 | PP: 0/1 | Block rank: 1 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.1 | PP: 0/1 | Block rank: 2 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.2 | PP: 0/1 | Block rank: 3 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.3 | PP: 0/1 | Block rank: 4 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.4 | PP: 0/1 | Block rank: 5 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.5 | PP: 0/1 | Block rank: 6 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.6 | PP: 0/1 | Block rank: 7 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.7 | PP: 0/1 | Block rank: 8 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.8 | PP: 0/1 | Block rank: 9 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.9 | PP: 0/1 | Block rank: 10 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.10 | PP: 0/1 | Block rank: 11 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.11 | PP: 0/1 | Block rank: 12 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.12 | PP: 0/1 | Block rank: 13 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.13 | PP: 0/1 | Block rank: 14 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.14 | PP: 0/1 | Block rank: 15 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.15 | PP: 0/1 | Block rank: 16 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.16 | PP: 0/1 | Block rank: 17 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.17 | PP: 0/1 | Block rank: 18 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.18 | PP: 0/1 | Block rank: 19 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.19 | PP: 0/1 | Block rank: 20 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.20 | PP: 0/1 | Block rank: 21 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.21 | PP: 0/1 | Block rank: 22 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.22 | PP: 0/1 | Block rank: 23 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.23 | PP: 0/1 | Block rank: 24 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.24 | PP: 0/1 | Block rank: 25 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.25 | PP: 0/1 | Block rank: 26 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.26 | PP: 0/1 | Block rank: 27 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.decoder.27 | PP: 0/1 | Block rank: 28 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.final_layer_norm | PP: 0/1 | Block rank: 29 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.lm_head | PP: 0/1 | Block rank: 30 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: model.cast_to_fp32 | PP: 0/1 | Block rank: 31 +01/07/2025 16:51:19 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: module_name: loss | PP: 0/1 | Block rank: 32 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/07/2025 16:51:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Total number of parameters: 3.57G (6817.34MiB) +01/07/2025 16:51:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Local number of parameters: 894M (1704.33MiB) +01/07/2025 16:51:23 [INFO|DP=0|PP=0|TP=1|ip-26-0-172-252]: Local number of parameters: 894M (1704.33MiB) +01/07/2025 16:51:23 [INFO|DP=0|PP=0|TP=3|ip-26-0-172-252]: Local number of parameters: 894M (1704.33MiB) +01/07/2025 16:51:23 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: [After model building] Memory usage: 1705.87MiB. Peak allocated: 5440.00MiB Peak reserved: 12610.00MiB +01/07/2025 16:51:23 [INFO|DP=0|PP=0|TP=2|ip-26-0-172-252]: Local number of parameters: 894M (1704.33MiB) +01/07/2025 16:51:23 [INFO|DP=0|PP=0|TP=1|ip-26-0-172-252]: [After model building] Memory usage: 1705.87MiB. Peak allocated: 5440.00MiB Peak reserved: 12610.00MiB +01/07/2025 16:51:23 [INFO|DP=0|PP=0|TP=3|ip-26-0-172-252]: [After model building] Memory usage: 1705.87MiB. Peak allocated: 5440.00MiB Peak reserved: 12610.00MiB +01/07/2025 16:51:23 [INFO|DP=0|PP=0|TP=2|ip-26-0-172-252]: [After model building] Memory usage: 1705.87MiB. Peak allocated: 5440.00MiB Peak reserved: 12610.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/07/2025 16:51:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: No checkpoint path provided. +01/07/2025 16:51:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Parametrizing model parameters using StandardParametrizator +01/07/2025 16:51:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: [Optimizer Building] Using LearningRateForSP as learning rate +01/07/2025 16:51:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +01/07/2025 16:51:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Using dummy data generator +01/07/2025 16:51:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: [Training Plan] There are 1 training stages +01/07/2025 16:51:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: [Stage Stable Training Stage] start from step 1 +01/07/2025 16:51:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: +01/07/2025 16:51:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: [Start training] datetime: 2025-01-07 16:51:27.727109 | mbs: 8 | grad_accum: 4 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +01/07/2025 16:51:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +01/07/2025 16:51:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Memory usage: 10227.54MiB. Peak allocated 10227.54MiB. Peak reserved: 21136.00MiB +01/07/2025 16:51:33 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Memory usage: 10697.12MiB. Peak allocated 46792.60MiB. Peak reserved: 62138.00MiB +01/07/2025 16:51:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 4.68K | tokens_per_sec: 224K | tokens_per_sec_per_gpu: 7K | global_batch_size: 256 | lm_loss: 12.4 | lr: 0.00015 | model_tflops_per_gpu: 180 | hardware_tflops_per_gpu: 180 | grad_norm: 1.46 | cuda_memory_allocated: 18.4G | cuda_max_memory_reserved: 65.2G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.3G | hd_free_memory_tb: 244G +01/07/2025 16:51:34 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Memory usage: 17514.54MiB. Peak allocated 17514.54MiB. Peak reserved: 62138.00MiB +01/07/2025 16:51:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Memory usage: 17514.54MiB. Peak allocated 53610.02MiB. Peak reserved: 66234.00MiB +01/07/2025 16:51:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 2.61K | tokens_per_sec: 402K | tokens_per_sec_per_gpu: 12.5K | global_batch_size: 256 | lm_loss: 12.4 | lr: 0.0003 | model_tflops_per_gpu: 322 | hardware_tflops_per_gpu: 322 | grad_norm: 1.46 | cuda_memory_allocated: 18.4G | cuda_max_memory_reserved: 69.5G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.3G | hd_free_memory_tb: 244G +01/07/2025 16:51:36 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Memory usage: 17514.54MiB. Peak allocated 17514.63MiB. Peak reserved: 66234.00MiB +01/07/2025 16:51:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Memory usage: 17514.54MiB. Peak allocated 53610.02MiB. Peak reserved: 66234.00MiB +01/07/2025 16:51:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 2.56K | tokens_per_sec: 409K | tokens_per_sec_per_gpu: 12.8K | global_batch_size: 256 | lm_loss: 12.4 | lr: 0.000296 | model_tflops_per_gpu: 328 | hardware_tflops_per_gpu: 328 | grad_norm: 1.44 | cuda_memory_allocated: 18.4G | cuda_max_memory_reserved: 69.5G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.3G | hd_free_memory_tb: 244G +01/07/2025 16:51:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Memory usage: 17514.54MiB. Peak allocated 17514.63MiB. Peak reserved: 66234.00MiB +01/07/2025 16:51:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Memory usage: 17514.54MiB. Peak allocated 53610.02MiB. Peak reserved: 66234.00MiB +01/07/2025 16:51:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: iteration: 4 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 2.56K | tokens_per_sec: 410K | tokens_per_sec_per_gpu: 12.8K | global_batch_size: 256 | lm_loss: 12.4 | lr: 0.000283 | model_tflops_per_gpu: 329 | hardware_tflops_per_gpu: 329 | grad_norm: 1.28 | cuda_memory_allocated: 18.4G | cuda_max_memory_reserved: 69.5G | hd_total_memory_tb: 312G | hd_used_memory_tb: 68.3G | hd_free_memory_tb: 244G +01/07/2025 16:51:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +01/07/2025 16:51:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: | -------- | ------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | ----------------------------------- | -------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +01/07/2025 16:51:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: | 14113419 | 3.57G_dp8_tp4_pp1_acc4_mbs8_seq4096_zero0_tpmodeRED_vocab131k | 4 | 4096 | 8 | 4 | 256 | 328.64 | 328.64 | 12802.99 | 169.29 | 99.66 | 100.68 | 424.42 | 240.04 | 242.41 | 17.10 | 64.68 | 8 | 1 | 4 | OneForwardOneBackwardPipelineEngine | REDUCE_SCATTER | True | 3072 | silu | 28 | 32 | 32 | 4096 | 131072 | True | torch.bfloat16 | 0 | 25 | True | 3.57G | 894M | +01/07/2025 16:51:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final2.csv +01/07/2025 16:51:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Acquired lock for CSV file: benchmark/results/bench_final2.csv +01/07/2025 16:51:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Successfully wrote to CSV file: benchmark/results/bench_final2.csv. Releasing lock... +01/07/2025 16:51:41 [INFO|DP=0|PP=0|TP=0|ip-26-0-172-252]: Throughput logging complete +[E TraceUtils.h:35] Store is down while updating #4635 with key NCCL_3_trace_start +[E TraceUtils.h:35] Store is down while updating #4636 with key NCCL_2_trace_start +[E TraceUtils.h:35] Store is down while updating #4638 with key NCCL_0_trace_end +[E TraceUtils.h:35] Store is down while updating #4637 with key NCCL_2_trace_start +[E TraceUtils.h:35] Store is down while updating #4635 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #4633 with key NCCL_0_trace_end +[E TraceUtils.h:35] Store is down while updating #4635 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #4636 with key NCCL_3_trace_start +[E TraceUtils.h:35] Store is down while updating #4635 with key NCCL_3_trace_start +[E TraceUtils.h:35] Store is down while updating #4636 with key NCCL_2_trace_start +[E TraceUtils.h:35] Store is down while updating #4636 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #4635 with key NCCL_3_trace_start +[E TraceUtils.h:35] Store is down while updating #4637 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #4637 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #4635 with key NCCL_2_trace_start +[E TraceUtils.h:35] Store is down while updating #4636 with key NCCL_1_trace_start +[2025-01-07 16:51:57,250] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 979347 closing signal SIGTERM +[2025-01-07 16:51:57,250] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 979349 closing signal SIGTERM +[2025-01-07 16:51:57,250] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 979350 closing signal SIGTERM +[2025-01-07 16:51:57,251] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 979351 closing signal SIGTERM +[2025-01-07 16:51:57,251] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 979352 closing signal SIGTERM +[2025-01-07 16:51:57,251] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 979353 closing signal SIGTERM +[2025-01-07 16:51:57,251] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 979354 closing signal SIGTERM +[2025-01-07 16:52:01,498] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: -11) local_rank: 1 (pid: 979348) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +======================================================== +/fsx/nouamane/projects/nanotron/run_train.py FAILED +-------------------------------------------------------- +Failures: + +-------------------------------------------------------- +Root Cause (first observed failure): +[0]: + time : 2025-01-07_16:51:57 + host : ip-26-0-173-7.ec2.internal + rank : 25 (local_rank: 1) + exitcode : -11 (pid: 979348) + error_file: + traceback : Signal 11 (SIGSEGV) received by PID 979348 +======================================================== +srun: error: ip-26-0-173-7: task 1: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=14113419.0 +slurmstepd: error: *** STEP 14113419.0 ON ip-26-0-172-252 CANCELLED AT 2025-01-07T16:52:01 *** +[2025-01-07 16:52:01,898] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67728 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67729 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67730 closing signal SIGTERM +[2025-01-07 16:52:01,898] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67731 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 67732 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206696 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206697 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206698 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206699 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206700 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206701 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206702 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 206703 closing signal SIGTERM +[2025-01-07 16:52:01,899] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-07 16:52:01,900] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 176418 closing signal SIGTERM +[2025-01-07 16:52:01,900] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 176419 closing signal SIGTERM +[2025-01-07 16:52:01,900] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 176420 closing signal SIGTERM +[2025-01-07 16:52:01,900] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 176421 closing signal SIGTERM +[2025-01-07 16:52:01,900] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 176422 closing signal SIGTERM +[2025-01-07 16:52:01,900] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 176423 closing signal SIGTERM +[2025-01-07 16:52:01,900] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 176424 closing signal SIGTERM +[2025-01-07 16:52:01,900] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 176425 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 67654 got signal: 15 +srun: error: ip-26-0-172-252: task 0: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 206626 got signal: 15 +[2025-01-07 16:52:07,056] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-173-202.ec2.internal_176348_0' has failed to send a keep-alive heartbeat to the rendezvous '14113419' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-173-121: task 2: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 176348 got signal: 15 +srun: error: ip-26-0-173-202: task 3: Exited with exit code 1 diff --git a/logs/14280992-bench_80G_dp2_tp8_pp16_acc128_mbs1_seq4096_zero0_tpmodeRED_vocab131k_gqa8.out b/logs/14280992-bench_80G_dp2_tp8_pp16_acc128_mbs1_seq4096_zero0_tpmodeRED_vocab131k_gqa8.out new file mode 100644 index 0000000000000000000000000000000000000000..88dd05556623b6f66650fbacb6d915146afdc293 --- /dev/null +++ b/logs/14280992-bench_80G_dp2_tp8_pp16_acc128_mbs1_seq4096_zero0_tpmodeRED_vocab131k_gqa8.out @@ -0,0 +1,3223 @@ ++ echo 'Running script: /var/spool/slurmd/job14280992/slurm_script' +Running script: /var/spool/slurmd/job14280992/slurm_script ++ '[' -z 14280992 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-161-[178,221],ip-26-0-162-[14,46,79,180],ip-26-0-165-[24,38,59,131,164,202,213],ip-26-0-166-[15,36,125],ip-26-0-167-[217,245],ip-26-0-168-[30,34,52,95,120,238],ip-26-0-169-[86,132,139,207,239,247],ip-26-0-170-[31,132]' ++ export 'NODELIST=ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-162-14 +ip-26-0-162-46 +ip-26-0-162-79 +ip-26-0-162-180 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132' ++ NODELIST='ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-162-14 +ip-26-0-162-46 +ip-26-0-162-79 +ip-26-0-162-180 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-161-[178,221],ip-26-0-162-[14,46,79,180],ip-26-0-165-[24,38,59,131,164,202,213],ip-26-0-166-[15,36,125],ip-26-0-167-[217,245],ip-26-0-168-[30,34,52,95,120,238],ip-26-0-169-[86,132,139,207,239,247],ip-26-0-170-[31,132]' ++ export MASTER_NODE=ip-26-0-161-178 ++ MASTER_NODE=ip-26-0-161-178 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=32 ++ NNODES=32 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=256 ++ WORLD_SIZE=256 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-43 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-43 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 52-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 52-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 52-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 52-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-161-178' +Master node: ip-26-0-161-178 ++ echo 'All nodes: ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-162-14 +ip-26-0-162-46 +ip-26-0-162-79 +ip-26-0-162-180 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132' +All nodes: ip-26-0-161-178 +ip-26-0-161-221 +ip-26-0-162-14 +ip-26-0-162-46 +ip-26-0-162-79 +ip-26-0-162-180 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-131 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-36 +ip-26-0-166-125 +ip-26-0-167-217 +ip-26-0-167-245 +ip-26-0-168-30 +ip-26-0-168-34 +ip-26-0-168-52 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-168-238 +ip-26-0-169-86 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-239 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 ++ echo 'World size: 256' +World size: 256 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=32 --nproc_per_node=8 --rdzv_id=14280992 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-161-178:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_80G_dp2_tp8_pp16_acc128_mbs1_seq4096_zero0_tpmodeRED_vocab131k_gqa8.yaml +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,463] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,464] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,464] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,466] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,466] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,466] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,466] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,466] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,466] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,467] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,468] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,468] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,471] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,473] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,473] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,473] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,473] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,473] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,474] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,474] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,474] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,474] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,474] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,475] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,475] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,475] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,475] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,475] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,486] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,491] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,463] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,463] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,463] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,463] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,462] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,506] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,464] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,464] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,464] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,464] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,466] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,466] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,466] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,466] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,464] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,464] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,464] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,464] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,467] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,467] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,467] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,467] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,508] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,471] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,471] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,471] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,471] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,468] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,468] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,468] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,468] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,470] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,468] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,468] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,468] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,468] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,486] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,486] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,486] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,486] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,491] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,491] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,491] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,491] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,506] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,506] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,506] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,506] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,508] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,508] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,508] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,508] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,553] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,578] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:17,553] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,553] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,553] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,553] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,579] torch.distributed.run: [WARNING] +[2025-01-08 11:46:17,579] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:17,579] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:17,579] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:21,115] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:21,114] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,176] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-08 11:46:21,177] torch.distributed.run: [WARNING] +[2025-01-08 11:46:21,177] torch.distributed.run: [WARNING] ***************************************** +[2025-01-08 11:46:21,177] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-08 11:46:21,177] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Config: +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Config(general=GeneralArgs(project='debug', +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: run='80G_dp2_tp8_pp16_acc128_mbs1_seq4096_zero0_tpmodeRED_vocab131k_gqa8', +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: seed=42, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: step=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: consumed_train_samples=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: ignore_sanity_checks=True), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: parallelism=ParallelismArgs(dp=2, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: pp=16, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: tp=8, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: pp_engine=, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: tp_mode=, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: tp_linear_async_communication=True, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: recompute_layer=False, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: tp_recompute_allgather=True, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: expert_parallel_size=1), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: eos_token_id=0, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: hidden_act='silu', +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: hidden_size=8192, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: initializer_range=0.02, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: intermediate_size=28672, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: is_llama_config=True, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: max_position_embeddings=4096, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: num_attention_heads=64, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: num_hidden_layers=80, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: num_key_value_heads=8, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: pad_token_id=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: pretraining_tp=1, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: rms_norm_eps=1e-05, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: rope_scaling=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: rope_theta=10000.0, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: rope_interleaved=False, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: tie_word_embeddings=False, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: use_cache=True, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: vocab_size=131072), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: init_method=RandomInit(std=0.02), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: dtype=torch.bfloat16, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: make_vocab_size_divisible_by=1, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: ddp_bucket_cap_mb=25), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: tokenizer_revision=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: tokenizer_max_length=None), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: checkpoint_interval=10000, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: save_initial_state=False, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: save_final_state=False, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: resume_checkpoint_path=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: checkpoints_path_is_shared_file_system=False), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: logging=LoggingArgs(log_level='info', +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: log_level_replica='info', +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: iteration_step_info_interval=1), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: tokens=TokensArgs(sequence_length=4096, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: train_steps=100, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: micro_batch_size=1, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: batch_accumulation_per_replica=128, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: val_check_interval=100, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: limit_val_batches=0, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: limit_test_batches=0), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: adam_beta1=0.9, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: adam_beta2=0.95, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: torch_adam_is_fused=True, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: name='adamW'), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: zero_stage=0, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: weight_decay=0.01, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: clip_grad=1.0, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: accumulate_grad_in_fp32=True, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: lr_warmup_steps=2, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: lr_warmup_style='linear', +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: lr_decay_style='cosine', +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: lr_decay_steps=13, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: lr_decay_starting_step=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: min_decay_lr=1e-05)), +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: start_training_step=1, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: data=DataArgs(dataset=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: seed=42, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: num_loading_workers=1))], +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: profiler=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: lighteval=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: s3_upload=None) +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Model Config: +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: LlamaConfig(bos_token_id=0, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: eos_token_id=0, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: hidden_act='silu', +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: hidden_size=8192, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: initializer_range=0.02, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: intermediate_size=28672, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: is_llama_config=True, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: max_position_embeddings=4096, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: num_attention_heads=64, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: num_hidden_layers=80, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: num_key_value_heads=8, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: pad_token_id=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: pretraining_tp=1, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: rms_norm_eps=1e-05, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: rope_scaling=None, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: rope_theta=10000.0, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: rope_interleaved=False, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: tie_word_embeddings=False, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: use_cache=True, +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: vocab_size=131072) +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Building model.. +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Initialize RoPE Theta = 10000.0 +01/08/2025 11:47:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Setting PP block ranks... +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.token_position_embeddings | PP: 0/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.0 | PP: 0/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.1 | PP: 0/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.2 | PP: 0/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.3 | PP: 0/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.4 | PP: 0/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.5 | PP: 0/16 | Block rank: 6 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.6 | PP: 1/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.7 | PP: 1/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.8 | PP: 1/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.9 | PP: 1/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.10 | PP: 1/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.11 | PP: 1/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.12 | PP: 2/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.13 | PP: 2/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.14 | PP: 2/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.15 | PP: 2/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.16 | PP: 2/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.17 | PP: 2/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.18 | PP: 3/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.19 | PP: 3/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.20 | PP: 3/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.21 | PP: 3/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.22 | PP: 3/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.23 | PP: 3/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.24 | PP: 4/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.25 | PP: 4/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.26 | PP: 4/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.27 | PP: 4/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.28 | PP: 4/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.29 | PP: 4/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.30 | PP: 5/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.31 | PP: 5/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.32 | PP: 5/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.33 | PP: 5/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.34 | PP: 5/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.35 | PP: 5/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.36 | PP: 6/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.37 | PP: 6/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.38 | PP: 6/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.39 | PP: 6/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.40 | PP: 6/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.41 | PP: 6/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.42 | PP: 7/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.43 | PP: 7/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.44 | PP: 7/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.45 | PP: 7/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.46 | PP: 7/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.47 | PP: 7/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.48 | PP: 8/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.49 | PP: 8/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.50 | PP: 8/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.51 | PP: 8/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.52 | PP: 8/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.53 | PP: 8/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.54 | PP: 9/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.55 | PP: 9/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.56 | PP: 9/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.57 | PP: 9/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.58 | PP: 9/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.59 | PP: 9/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.60 | PP: 10/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.61 | PP: 10/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.62 | PP: 10/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.63 | PP: 10/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.64 | PP: 10/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.65 | PP: 10/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.66 | PP: 11/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.67 | PP: 11/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.68 | PP: 11/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.69 | PP: 11/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.70 | PP: 11/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.71 | PP: 11/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.72 | PP: 12/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.73 | PP: 12/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.74 | PP: 12/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.75 | PP: 12/16 | Block rank: 3 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.76 | PP: 12/16 | Block rank: 4 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.77 | PP: 12/16 | Block rank: 5 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.78 | PP: 13/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.decoder.79 | PP: 14/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.final_layer_norm | PP: 15/16 | Block rank: 0 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.lm_head | PP: 15/16 | Block rank: 1 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: model.cast_to_fp32 | PP: 15/16 | Block rank: 2 +01/08/2025 11:47:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: module_name: loss | PP: 15/16 | Block rank: 3 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Total number of parameters: 70.6G (134676.12MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Local number of parameters: 776M (1480.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=0|ip-26-0-168-30]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=0|ip-26-0-165-202]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=0|ip-26-0-167-217]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=0|ip-26-0-169-247]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=0|ip-26-0-168-52]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=0|ip-26-0-169-132]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=0|ip-26-0-165-59]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=0|ip-26-0-165-24]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: [After model building] Memory usage: 1480.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=0|ip-26-0-162-14]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=0|ip-26-0-168-120]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=0|ip-26-0-162-46]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=0|ip-26-0-170-132]: Local number of parameters: 134M (256.02MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=0|ip-26-0-169-207]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=0|ip-26-0-165-202]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=0|ip-26-0-168-30]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=0|ip-26-0-169-247]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=0|ip-26-0-168-52]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=0|ip-26-0-169-132]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=0|ip-26-0-165-24]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=0|ip-26-0-165-59]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 11542.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=0|ip-26-0-167-217]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=0|ip-26-0-162-14]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=0|ip-26-0-168-120]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=0|ip-26-0-162-46]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=0|ip-26-0-170-132]: [After model building] Memory usage: 256.03MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=0|ip-26-0-169-207]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=0|ip-26-0-165-131]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=0|ip-26-0-166-15]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=0|ip-26-0-165-131]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=0|ip-26-0-166-15]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=6|ip-26-0-161-178]: Local number of parameters: 776M (1480.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=6|ip-26-0-161-178]: [After model building] Memory usage: 1480.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=7|ip-26-0-161-178]: Local number of parameters: 776M (1480.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=7|ip-26-0-161-178]: [After model building] Memory usage: 1480.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=4|ip-26-0-161-178]: Local number of parameters: 776M (1480.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=3|ip-26-0-161-178]: Local number of parameters: 776M (1480.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=4|ip-26-0-161-178]: [After model building] Memory usage: 1480.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=5|ip-26-0-161-178]: Local number of parameters: 776M (1480.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=2|ip-26-0-161-178]: Local number of parameters: 776M (1480.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=3|ip-26-0-161-178]: [After model building] Memory usage: 1480.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=5|ip-26-0-161-178]: [After model building] Memory usage: 1480.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=2|ip-26-0-161-178]: [After model building] Memory usage: 1480.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=1|ip-26-0-169-132]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=1|ip-26-0-162-14]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=1|ip-26-0-168-52]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=1|ip-26-0-162-46]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=1|ip-26-0-167-217]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=1|ip-26-0-169-247]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=1|ip-26-0-165-131]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=1|ip-26-0-168-30]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=1|ip-26-0-168-120]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=1|ip-26-0-161-178]: Local number of parameters: 776M (1480.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=1|ip-26-0-165-202]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=1|ip-26-0-165-59]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=1|ip-26-0-169-132]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=1|ip-26-0-162-14]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=1|ip-26-0-162-46]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=1|ip-26-0-167-217]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=1|ip-26-0-168-52]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=1|ip-26-0-168-120]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=1|ip-26-0-165-131]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=1|ip-26-0-168-30]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=1|ip-26-0-169-247]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=1|ip-26-0-165-202]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=1|ip-26-0-165-59]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=0|TP=1|ip-26-0-161-178]: [After model building] Memory usage: 1480.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=7|ip-26-0-167-217]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=7|ip-26-0-165-202]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=7|ip-26-0-169-247]: Local number of parameters: 107M (204.03MiB) +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=7|ip-26-0-162-46]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=7|ip-26-0-169-132]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=7|ip-26-0-162-14]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=7|ip-26-0-169-207]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=7|ip-26-0-165-131]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=7|ip-26-0-168-120]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=7|ip-26-0-168-30]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=7|ip-26-0-165-59]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=7|ip-26-0-168-52]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=7|ip-26-0-165-24]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=7|ip-26-0-166-15]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=7|ip-26-0-170-132]: Local number of parameters: 134M (256.02MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=7|ip-26-0-167-217]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=6|ip-26-0-162-14]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=6|ip-26-0-165-202]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=6|ip-26-0-169-132]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=6|ip-26-0-167-217]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=6|ip-26-0-165-131]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=6|ip-26-0-169-247]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=7|ip-26-0-162-46]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=6|ip-26-0-169-207]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=7|ip-26-0-169-207]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=7|ip-26-0-165-202]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=7|ip-26-0-169-132]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=7|ip-26-0-168-120]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=7|ip-26-0-162-14]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=7|ip-26-0-165-131]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=7|ip-26-0-169-247]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=6|ip-26-0-162-46]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=6|ip-26-0-168-30]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=6|ip-26-0-168-120]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=7|ip-26-0-165-59]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=3|ip-26-0-167-217]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=7|ip-26-0-168-52]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=7|ip-26-0-166-15]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=7|ip-26-0-170-132]: [After model building] Memory usage: 256.03MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=7|ip-26-0-168-30]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=7|ip-26-0-165-24]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=3|ip-26-0-162-46]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=3|ip-26-0-162-14]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=6|ip-26-0-170-132]: Local number of parameters: 134M (256.02MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=3|ip-26-0-166-15]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=3|ip-26-0-169-247]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=6|ip-26-0-162-14]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=3|ip-26-0-169-132]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=6|ip-26-0-168-52]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=3|ip-26-0-165-202]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=3|ip-26-0-168-30]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=6|ip-26-0-165-131]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=6|ip-26-0-169-132]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=5|ip-26-0-167-217]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=6|ip-26-0-167-217]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=3|ip-26-0-168-120]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=3|ip-26-0-169-207]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=3|ip-26-0-168-52]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=5|ip-26-0-162-14]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=6|ip-26-0-169-247]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=6|ip-26-0-165-202]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=3|ip-26-0-170-132]: Local number of parameters: 134M (256.02MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=6|ip-26-0-169-207]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=3|ip-26-0-165-24]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=5|ip-26-0-166-15]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=6|ip-26-0-168-30]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=5|ip-26-0-168-30]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=3|ip-26-0-165-131]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=3|ip-26-0-165-59]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=5|ip-26-0-165-59]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=5|ip-26-0-165-202]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=5|ip-26-0-169-132]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=2|ip-26-0-169-132]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=2|ip-26-0-167-217]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=6|ip-26-0-162-46]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=5|ip-26-0-162-46]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=5|ip-26-0-165-24]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=2|ip-26-0-169-207]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=5|ip-26-0-168-52]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=5|ip-26-0-170-132]: Local number of parameters: 134M (256.02MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=2|ip-26-0-162-14]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=2|ip-26-0-165-59]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=6|ip-26-0-168-120]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=3|ip-26-0-167-217]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=2|ip-26-0-162-46]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=3|ip-26-0-162-46]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=2|ip-26-0-165-24]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=5|ip-26-0-169-207]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=3|ip-26-0-162-14]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=5|ip-26-0-165-131]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=5|ip-26-0-169-247]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=2|ip-26-0-165-202]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=6|ip-26-0-168-52]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=6|ip-26-0-170-132]: [After model building] Memory usage: 256.03MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=2|ip-26-0-166-15]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=3|ip-26-0-166-15]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=3|ip-26-0-169-132]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=2|ip-26-0-169-247]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=3|ip-26-0-169-247]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=3|ip-26-0-165-202]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=3|ip-26-0-168-52]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=5|ip-26-0-167-217]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=3|ip-26-0-169-207]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=2|ip-26-0-170-132]: Local number of parameters: 134M (256.02MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=5|ip-26-0-162-14]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=2|ip-26-0-168-30]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=3|ip-26-0-168-30]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=5|ip-26-0-168-120]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=3|ip-26-0-170-132]: [After model building] Memory usage: 256.03MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=5|ip-26-0-165-202]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=5|ip-26-0-169-132]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=2|ip-26-0-169-132]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=1|ip-26-0-165-24]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=3|ip-26-0-168-120]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=2|ip-26-0-169-207]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=2|ip-26-0-168-52]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=5|ip-26-0-170-132]: [After model building] Memory usage: 256.03MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=5|ip-26-0-165-59]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 11542.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=5|ip-26-0-168-30]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=2|ip-26-0-162-14]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=2|ip-26-0-165-131]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=5|ip-26-0-166-15]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=2|ip-26-0-167-217]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=3|ip-26-0-165-24]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=5|ip-26-0-165-24]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=5|ip-26-0-168-52]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=5|ip-26-0-162-46]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=5|ip-26-0-169-207]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=2|ip-26-0-165-59]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=3|ip-26-0-165-59]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=2|ip-26-0-168-120]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=2|ip-26-0-165-24]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=3|ip-26-0-165-131]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=5|ip-26-0-165-131]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=2|ip-26-0-162-46]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=2|ip-26-0-165-202]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=5|ip-26-0-169-247]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=2|ip-26-0-169-247]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=6|ip-26-0-165-24]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=6|ip-26-0-166-15]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=1|ip-26-0-166-15]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=2|ip-26-0-166-15]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=5|ip-26-0-168-120]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=2|ip-26-0-170-132]: [After model building] Memory usage: 256.03MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=6|ip-26-0-165-59]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=2|ip-26-0-168-30]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=1|ip-26-0-165-24]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=1|ip-26-0-169-207]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=2|ip-26-0-165-131]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=2|ip-26-0-168-52]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=2|ip-26-0-168-120]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=1|ip-26-0-170-132]: Local number of parameters: 134M (256.02MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=6|ip-26-0-165-59]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=6|ip-26-0-165-24]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=1|ip-26-0-166-15]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=6|ip-26-0-166-15]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=1|ip-26-0-169-207]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=1|ip-26-0-170-132]: [After model building] Memory usage: 256.03MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=4|ip-26-0-167-217]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=4|ip-26-0-165-202]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=4|ip-26-0-169-132]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=4|ip-26-0-162-14]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=4|ip-26-0-165-131]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=4|ip-26-0-169-247]: Local number of parameters: 107M (204.03MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=4|ip-26-0-170-132]: Local number of parameters: 134M (256.02MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=4|ip-26-0-165-59]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=4|ip-26-0-162-46]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=4|ip-26-0-168-52]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=4|ip-26-0-168-30]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=4|ip-26-0-166-15]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=4|ip-26-0-165-24]: Local number of parameters: 642M (1224.19MiB) +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=4|ip-26-0-169-207]: Local number of parameters: 107M (204.03MiB) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=8|TP=4|ip-26-0-167-217]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=4|TP=4|ip-26-0-165-202]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=4|ip-26-0-168-120]: Local number of parameters: 642M (1224.19MiB) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=1|TP=4|ip-26-0-162-14]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=3|TP=4|ip-26-0-165-131]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=12|TP=4|ip-26-0-169-132]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=15|TP=4|ip-26-0-170-132]: [After model building] Memory usage: 256.03MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=11|TP=4|ip-26-0-168-52]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=6|TP=4|ip-26-0-165-59]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 11542.00MiB +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=14|TP=4|ip-26-0-169-247]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=2|TP=4|ip-26-0-162-46]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=10|TP=4|ip-26-0-168-30]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=7|TP=4|ip-26-0-166-15]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +01/08/2025 11:47:29 [INFO|DP=0|PP=5|TP=4|ip-26-0-165-24]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=13|TP=4|ip-26-0-169-207]: [After model building] Memory usage: 204.04MiB. Peak allocated: 5384.00MiB Peak reserved: 10518.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:29 [INFO|DP=0|PP=9|TP=4|ip-26-0-168-120]: [After model building] Memory usage: 1224.20MiB. Peak allocated: 5384.00MiB Peak reserved: 9494.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/08/2025 11:47:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: No checkpoint path provided. +01/08/2025 11:47:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Parametrizing model parameters using StandardParametrizator +01/08/2025 11:47:30 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: [Optimizer Building] Using LearningRateForSP as learning rate +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +01/08/2025 11:47:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +01/08/2025 11:47:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Using dummy data generator +01/08/2025 11:47:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: [Training Plan] There are 1 training stages +01/08/2025 11:47:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: [Stage Stable Training Stage] start from step 1 +01/08/2025 11:47:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: +01/08/2025 11:47:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: [Start training] datetime: 2025-01-08 11:47:31.123017 | mbs: 1 | grad_accum: 128 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +01/08/2025 11:47:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +01/08/2025 11:47:31 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Memory usage: 8881.14MiB. Peak allocated 8881.14MiB. Peak reserved: 16900.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +01/08/2025 11:48:46 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Memory usage: 9079.18MiB. Peak allocated 25364.54MiB. Peak reserved: 29708.00MiB +01/08/2025 11:48:56 [INFO|DP=0|PP=15|TP=0|ip-26-0-170-132]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 83.7K | tokens_per_sec: 12.5K | tokens_per_sec_per_gpu: 48.9 | global_batch_size: 256 | lm_loss: 13.4 | lr: 0.00015 | model_tflops_per_gpu: 22 | hardware_tflops_per_gpu: 22 | grad_norm: 10.2 | cuda_memory_allocated: 2.82G | cuda_max_memory_reserved: 5.37G | hd_total_memory_tb: 312G | hd_used_memory_tb: 94.6G | hd_free_memory_tb: 218G +01/08/2025 11:48:56 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Memory usage: 14999.95MiB. Peak allocated 14999.95MiB. Peak reserved: 33580.00MiB +01/08/2025 11:49:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Memory usage: 14999.95MiB. Peak allocated 31285.31MiB. Peak reserved: 34204.00MiB +01/08/2025 11:49:18 [INFO|DP=0|PP=15|TP=0|ip-26-0-170-132]: iteration: 2 / 100 | consumed_tokens: 2.1M | elapsed_time_per_iteration_ms: 22.1K | tokens_per_sec: 47.5K | tokens_per_sec_per_gpu: 186 | global_batch_size: 256 | lm_loss: 13.4 | lr: 0.0003 | model_tflops_per_gpu: 83.4 | hardware_tflops_per_gpu: 83.4 | grad_norm: 10.2 | cuda_memory_allocated: 2.82G | cuda_max_memory_reserved: 5.37G | hd_total_memory_tb: 312G | hd_used_memory_tb: 94.6G | hd_free_memory_tb: 218G +01/08/2025 11:49:18 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Memory usage: 14999.95MiB. Peak allocated 14999.97MiB. Peak reserved: 34204.00MiB +01/08/2025 11:49:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Memory usage: 14999.95MiB. Peak allocated 31285.31MiB. Peak reserved: 34204.00MiB +01/08/2025 11:49:40 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Memory usage: 14999.95MiB. Peak allocated 14999.97MiB. Peak reserved: 34204.00MiB +01/08/2025 11:49:40 [INFO|DP=0|PP=15|TP=0|ip-26-0-170-132]: iteration: 3 / 100 | consumed_tokens: 3.15M | elapsed_time_per_iteration_ms: 21.4K | tokens_per_sec: 49K | tokens_per_sec_per_gpu: 191 | global_batch_size: 256 | lm_loss: 13.4 | lr: 0.000296 | model_tflops_per_gpu: 86 | hardware_tflops_per_gpu: 86 | grad_norm: 10 | cuda_memory_allocated: 2.82G | cuda_max_memory_reserved: 5.37G | hd_total_memory_tb: 312G | hd_used_memory_tb: 94.6G | hd_free_memory_tb: 218G +01/08/2025 11:50:00 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Memory usage: 14999.95MiB. Peak allocated 31285.31MiB. Peak reserved: 34204.00MiB +01/08/2025 11:50:01 [INFO|DP=0|PP=15|TP=0|ip-26-0-170-132]: iteration: 4 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 21K | tokens_per_sec: 50K | tokens_per_sec_per_gpu: 195 | global_batch_size: 256 | lm_loss: 13.4 | lr: 0.000283 | model_tflops_per_gpu: 87.8 | hardware_tflops_per_gpu: 87.8 | grad_norm: 9.65 | cuda_memory_allocated: 2.82G | cuda_max_memory_reserved: 5.37G | hd_total_memory_tb: 312G | hd_used_memory_tb: 94.6G | hd_free_memory_tb: 218G +01/08/2025 11:50:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | Total Params | Local Params | +01/08/2025 11:50:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: | -------- | ------------------------------------------------------------------- | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | -- | -- | -- | ----------------------------------- | -------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | ------------ | ------------ | +01/08/2025 11:50:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: | 14280992 | 80G_dp2_tp8_pp16_acc128_mbs1_seq4096_zero0_tpmodeRED_vocab131k_gqa8 | 32 | 4096 | 1 | 128 | 256 | 87.81 | 87.81 | 195.40 | 26.62 | 13.02 | 13.09 | 414.81 | 237.47 | 237.17 | 14.65 | 33.40 | 2 | 16 | 8 | OneForwardOneBackwardPipelineEngine | REDUCE_SCATTER | True | 8192 | silu | 80 | 64 | 8 | 4096 | 131072 | False | torch.bfloat16 | 0 | 25 | True | 70.6G | 776M | +01/08/2025 11:50:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Attempting to write benchmark results to CSV file: benchmark/results/bench_final2.csv +01/08/2025 11:50:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Acquired lock for CSV file: benchmark/results/bench_final2.csv +01/08/2025 11:50:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Successfully wrote to CSV file: benchmark/results/bench_final2.csv. Releasing lock... +01/08/2025 11:50:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-161-178]: Throughput logging complete +[2025-01-08 11:50:28,624] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 1617375) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2025-01-08 11:50:28,634] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 0 (pid: 195280) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2025-01-08 11:50:28,637] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915359 closing signal SIGTERM +[2025-01-08 11:50:28,638] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915360 closing signal SIGTERM +[2025-01-08 11:50:28,638] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915361 closing signal SIGTERM +[2025-01-08 11:50:28,638] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915362 closing signal SIGTERM +[2025-01-08 11:50:28,638] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915363 closing signal SIGTERM +[2025-01-08 11:50:28,638] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915364 closing signal SIGTERM +[2025-01-08 11:50:28,638] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915365 closing signal SIGTERM +[2025-01-08 11:50:28,640] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579081 closing signal SIGTERM +[2025-01-08 11:50:28,640] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579082 closing signal SIGTERM +[2025-01-08 11:50:28,640] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579083 closing signal SIGTERM +[2025-01-08 11:50:28,640] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579084 closing signal SIGTERM +[2025-01-08 11:50:28,640] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579085 closing signal SIGTERM +[2025-01-08 11:50:28,640] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579086 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2025-01-08_11:50:28 + host : ip-26-0-170-31.ec2.internal + rank : 249 (local_rank: 1) + exitcode : 1 (pid: 1617376) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2025-01-08_11:50:28 + host : ip-26-0-170-31.ec2.internal + rank : 250 (local_rank: 2) + exitcode : 1 (pid: 1617377) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2025-01-08_11:50:28 + host : ip-26-0-170-31.ec2.internal + rank : 251 (local_rank: 3) + exitcode : 1 (pid: 1617378) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2025-01-08_11:50:28 + host : ip-26-0-170-31.ec2.internal + rank : 252 (local_rank: 4) + exitcode : 1 (pid: 1617379) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2025-01-08_11:50:28 + host : ip-26-0-170-31.ec2.internal + rank : 253 (local_rank: 5) + exitcode : 1 (pid: 1617380) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2025-01-08_11:50:28 + host : ip-26-0-170-31.ec2.internal + rank : 254 (local_rank: 6) + exitcode : 1 (pid: 1617381) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2025-01-08_11:50:28 + host : ip-26-0-170-31.ec2.internal + rank : 255 (local_rank: 7) + exitcode : 1 (pid: 1617382) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2025-01-08_11:50:28 + host : ip-26-0-170-31.ec2.internal + rank : 248 (local_rank: 0) + exitcode : 1 (pid: 1617375) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: +[1]: + time : 2025-01-08_11:50:28 + host : ip-26-0-161-178.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 195281) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[2]: + time : 2025-01-08_11:50:28 + host : ip-26-0-161-178.ec2.internal + rank : 2 (local_rank: 2) + exitcode : 1 (pid: 195282) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[3]: + time : 2025-01-08_11:50:28 + host : ip-26-0-161-178.ec2.internal + rank : 3 (local_rank: 3) + exitcode : 1 (pid: 195283) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[4]: + time : 2025-01-08_11:50:28 + host : ip-26-0-161-178.ec2.internal + rank : 4 (local_rank: 4) + exitcode : 1 (pid: 195284) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[5]: + time : 2025-01-08_11:50:28 + host : ip-26-0-161-178.ec2.internal + rank : 5 (local_rank: 5) + exitcode : 1 (pid: 195285) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[6]: + time : 2025-01-08_11:50:28 + host : ip-26-0-161-178.ec2.internal + rank : 6 (local_rank: 6) + exitcode : 1 (pid: 195286) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +[7]: + time : 2025-01-08_11:50:28 + host : ip-26-0-161-178.ec2.internal + rank : 7 (local_rank: 7) + exitcode : 1 (pid: 195287) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2025-01-08_11:50:28 + host : ip-26-0-161-178.ec2.internal + rank : 0 (local_rank: 0) + exitcode : 1 (pid: 195280) + error_file: + traceback : To enable traceback see: https://pytorch.org/docs/stable/elastic/errors.html +============================================================ +srun: error: ip-26-0-161-178: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=14280992.0 +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579081 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579082 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579083 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 303824 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 303825 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579084 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 192949 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 303826 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 303827 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579085 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 303828 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 303829 closing signal SIGTERM +[2025-01-08 11:50:28,941] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 192950 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 579086 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 303830 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 344904 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 192951 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 303831 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 344905 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 192952 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 192953 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 344906 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 344907 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 344908 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 192954 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 510296 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 510297 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 344909 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106162 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 344910 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 192955 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 510298 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106163 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 510299 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106164 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 444123 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106165 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 510300 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106166 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 344911 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 192956 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 219307 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 510301 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106167 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 219308 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 510302 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 444124 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106168 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 219309 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 510303 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249090 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 444125 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 106169 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 219310 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249091 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83313 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249092 closing signal SIGTERM +[2025-01-08 11:50:28,942] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83314 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 444126 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 306307 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 219311 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249093 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 306308 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 219312 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83315 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249094 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 219313 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83316 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 444127 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 416638 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 306309 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205400 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83317 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249095 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83318 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 306310 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 219314 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 33471 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205401 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83319 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249096 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 306311 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83320 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 249097 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 444128 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 416639 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 306312 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 33472 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205402 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205403 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 416640 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 306313 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 33473 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205404 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915359 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 444129 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205405 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 444130 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205406 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421215 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22314 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 205407 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 211876 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 416641 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 306314 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 33474 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915360 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421216 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 211877 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22315 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22316 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915361 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 416642 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 33475 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915362 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 211878 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421217 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74941 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22317 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915363 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421218 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74942 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22318 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915364 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 211879 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 416643 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421219 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74943 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79262 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22319 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 33476 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1915365 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74944 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22320 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52956 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 211880 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 250359 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421220 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74945 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 22321 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 416644 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74946 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1126344 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 33477 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52957 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 416645 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 250360 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74947 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79263 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 33478 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 250361 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 74948 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1126345 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52958 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 211881 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 250362 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421221 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79264 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78929 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52959 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 250363 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 421222 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78930 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52960 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 250364 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79265 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1126346 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124754 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78931 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1126347 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 211882 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 250365 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78932 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52961 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 211883 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 250366 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1126348 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52962 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918006 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79266 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124755 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78933 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1126349 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78934 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918007 closing signal SIGTERM +[2025-01-08 11:50:28,945] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 52963 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124756 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79267 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78935 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1126350 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124757 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918008 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1126351 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 78936 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79268 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 79269 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124758 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918009 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58250 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124759 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918010 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918011 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918012 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58251 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124760 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 124761 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 918013 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58252 closing signal SIGTERM +[2025-01-08 11:50:28,943] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58253 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58254 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58255 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212753 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212754 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58256 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 58257 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212755 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212756 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212757 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212758 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212759 closing signal SIGTERM +[2025-01-08 11:50:28,945] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 212760 closing signal SIGTERM +[2025-01-08 11:50:28,944] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,945] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 329156 closing signal SIGTERM +[2025-01-08 11:50:28,945] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 329157 closing signal SIGTERM +[2025-01-08 11:50:28,945] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 329158 closing signal SIGTERM +[2025-01-08 11:50:28,945] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 329159 closing signal SIGTERM +[2025-01-08 11:50:28,945] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 329160 closing signal SIGTERM +[2025-01-08 11:50:28,945] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 329161 closing signal SIGTERM +[2025-01-08 11:50:28,945] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 329162 closing signal SIGTERM +[2025-01-08 11:50:28,945] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 329163 closing signal SIGTERM +[2025-01-08 11:50:28,947] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-08 11:50:28,947] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 947241 closing signal SIGTERM +[2025-01-08 11:50:28,948] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 947242 closing signal SIGTERM +[2025-01-08 11:50:28,948] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 947243 closing signal SIGTERM +[2025-01-08 11:50:28,948] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 947244 closing signal SIGTERM +[2025-01-08 11:50:28,948] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 947245 closing signal SIGTERM +[2025-01-08 11:50:28,948] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 947246 closing signal SIGTERM +[2025-01-08 11:50:28,948] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 947247 closing signal SIGTERM +[2025-01-08 11:50:28,948] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 947248 closing signal SIGTERM +srun: error: ip-26-0-170-31: task 30: Terminated +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 579003 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 878, in _invoke_run + run_result = self._monitor_workers(self._worker_group) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 307, in _monitor_workers + result = self._pcontext.wait(0) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 288, in wait + return self._poll() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 668, in _poll + self.close() # terminate all running procs + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1915179 got signal: 15 +srun: error: ip-26-0-161-221: task 1: Exited with exit code 1 +[2025-01-08 11:50:32,696] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-95.ec2.internal_124677_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,721] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-120.ec2.internal_329079_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,728] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-52.ec2.internal_52879_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,750] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-15.ec2.internal_306229_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,760] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-238.ec2.internal_250282_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,775] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-162-79.ec2.internal_303747_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,779] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-34.ec2.internal_212677_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-170-132: task 31: Exited with exit code 1 +[2025-01-08 11:50:32,810] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-245.ec2.internal_510217_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,819] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-59.ec2.internal_416561_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,845] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-162-46.ec2.internal_344827_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,859] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-164.ec2.internal_211798_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,862] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-24.ec2.internal_192870_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,869] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-132.ec2.internal_421139_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,881] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-38.ec2.internal_219231_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,896] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-139.ec2.internal_917930_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,913] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-162-180.ec2.internal_106085_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,919] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-36.ec2.internal_22237_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,923] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-131.ec2.internal_249012_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,945] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-213.ec2.internal_444046_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,957] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-217.ec2.internal_74864_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,965] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-30.ec2.internal_79184_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,973] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-166-125.ec2.internal_205321_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:32,980] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-202.ec2.internal_33395_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:33,028] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-86.ec2.internal_947164_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:33,340] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-207.ec2.internal_58174_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:33,492] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-247.ec2.internal_78822_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +[2025-01-08 11:50:33,515] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-169-239.ec2.internal_1126266_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper +[2025-01-08 11:50:33,548] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-162-14.ec2.internal_83236_0' has failed to send a keep-alive heartbeat to the rendezvous '14280992' due to an error of type RendezvousConnectionError. + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 205321 got signal: 15 +srun: error: ip-26-0-166-125: task 15: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 947164 got signal: 15 +srun: error: ip-26-0-169-86: task 24: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 249012 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 83236 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 212677 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 192870 got signal: 15 +srun: error: ip-26-0-165-131: task 9: Exited with exit code 1 +srun: error: ip-26-0-162-14: task 2: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 219231 got signal: 15 +srun: error: ip-26-0-168-34: task 19: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 52879 got signal: 15 +srun: error: ip-26-0-165-24: task 6: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 78822 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 306229 got signal: 15 +srun: error: ip-26-0-165-38: task 7: Exited with exit code 1 +srun: error: ip-26-0-168-52: task 20: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 329079 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 22237 got signal: 15 +srun: error: ip-26-0-169-247: task 29: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 58174 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 250282 got signal: 15 +srun: error: ip-26-0-166-15: task 13: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1126266 got signal: 15 +srun: error: ip-26-0-166-36: task 14: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 510217 got signal: 15 +srun: error: ip-26-0-169-207: task 27: Exited with exit code 1 +srun: error: ip-26-0-168-238: task 23: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 444046 got signal: 15 +srun: error: ip-26-0-168-120: task 22: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 33395 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 421139 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 79184 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 917930 got signal: 15 + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 106085 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 344827 got signal: 15 +srun: error: ip-26-0-169-239: task 28: Exited with exit code 1 +srun: error: ip-26-0-167-245: task 17: Exited with exit code 1 +srun: error: ip-26-0-165-213: task 12: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 211798 got signal: 15 +srun: error: ip-26-0-169-139: task 26: Exited with exit code 1 +srun: error: ip-26-0-165-202: task 11: Exited with exit code 1 +srun: error: ip-26-0-168-30: task 18: Exited with exit code 1 +srun: error: ip-26-0-169-132: task 25: Exited with exit code 1 +srun: error: ip-26-0-162-46: task 3: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 416561 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run +srun: error: ip-26-0-162-180: task 5: Exited with exit code 1 + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 303747 got signal: 15 +srun: error: ip-26-0-165-164: task 10: Exited with exit code 1 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 124677 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 74864 got signal: 15 +srun: error: ip-26-0-165-59: task 8: Exited with exit code 1 +srun: error: ip-26-0-162-79: task 4: Exited with exit code 1 +srun: error: ip-26-0-168-95: task 21: Exited with exit code 1 +srun: error: ip-26-0-167-217: task 16: Exited with exit code 1 +srun: Force Terminated StepId=14280992.0 diff --git a/logs/14471696-bench_80G_dp1_tp4_pp16_acc256_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out b/logs/14471696-bench_80G_dp1_tp4_pp16_acc256_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out new file mode 100644 index 0000000000000000000000000000000000000000..ef56000f71fbe98eb7f0ff67cbca94a0d016e156 --- /dev/null +++ b/logs/14471696-bench_80G_dp1_tp4_pp16_acc256_mbs1_seq4096_zero0_tpmodeRED_vocab131k.out @@ -0,0 +1,1584 @@ ++ '[' -z 14471696 ']' ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.local/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh"' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=2 ++++ CONDA_SHLVL=2 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_1=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ . /fsx/nouamane/miniconda/etc/conda/activate.d/libxml2_activate.sh +++++ test -n '' +++++ xml_catalog_files_libxml2= +++++ XML_CATALOG_FILES= +++++ conda_catalog_files= +++++ ifs_libxml2=' +' +++++ IFS=' ' +++++ rem=/fsx/nouamane/miniconda +++++ for pre in ${rem} +++++ test '' = /fsx/nouamane/miniconda +++++ conda_catalog_files=/fsx/nouamane/miniconda +++++ rem= +++++ IFS=' +' +++++ conda_catalog_files='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/etc/xml/catalog file:///etc/xml/catalog' +++++ unset conda_catalog_files ifs_libxml2 rem +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 +Error while loading conda entry point: conda-libmamba-solver (module 'libmambapy' has no attribute 'QueryFormat') ++ ask_conda='. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval '. "/fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ . /fsx/nouamane/miniconda/etc/conda/deactivate.d/libxml2_deactivate.sh ++++ test -n '' ++++ unset XML_CATALOG_FILES ++++ unset xml_catalog_files_libxml2 +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=3 +++ CONDA_SHLVL=3 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ CONDA_PREFIX_2=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/316e524257c2ea23b755332b0a72c50cf23e1b00/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.local/bin:/admin/home/nouamane/.fzf/bin ++ '[' -z '' ']' +++ scontrol show hostnames 'ip-26-0-163-[134,158],ip-26-0-164-0,ip-26-0-165-164,ip-26-0-167-[111,175,217],ip-26-0-168-30' ++ export 'NODELIST=ip-26-0-163-134 +ip-26-0-163-158 +ip-26-0-164-0 +ip-26-0-165-164 +ip-26-0-167-111 +ip-26-0-167-175 +ip-26-0-167-217 +ip-26-0-168-30' ++ NODELIST='ip-26-0-163-134 +ip-26-0-163-158 +ip-26-0-164-0 +ip-26-0-165-164 +ip-26-0-167-111 +ip-26-0-167-175 +ip-26-0-167-217 +ip-26-0-168-30' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-163-[134,158],ip-26-0-164-0,ip-26-0-165-164,ip-26-0-167-[111,175,217],ip-26-0-168-30' ++ export MASTER_NODE=ip-26-0-163-134 ++ MASTER_NODE=ip-26-0-163-134 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=8 ++ NNODES=8 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=64 ++ WORLD_SIZE=64 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NCCL_DEBUG=WARN ++ NCCL_DEBUG=WARN ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ export WANDB_MODE=disabled ++ WANDB_MODE=disabled ++ export TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ TORCH_NCCL_ASYNC_ERROR_HANDLING=1 ++ export TORCH_DISTRIBUTED_DEBUG=DETAIL ++ TORCH_DISTRIBUTED_DEBUG=DETAIL ++ '[' -z '' ']' ++ echo '=== GPU Topology ===' +=== GPU Topology === ++ nvidia-smi topo -m + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 CPU Affinity NUMA Affinity GPU NUMA ID +GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 0-47 0 N/A +GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 0-47 0 N/A +GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 48-95 1 N/A +GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 48-95 1 N/A +GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 48-95 1 N/A +GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X 48-95 1 N/A + +Legend: + + X = Self + SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI) + NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node + PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU) + PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge) + PIX = Connection traversing at most a single PCIe bridge + NV# = Connection traversing a bonded set of # NVLinks ++ echo ================== +================== ++ export SRUN_ALLOC_ARGS= ++ SRUN_ALLOC_ARGS= ++ echo 'Master node: ip-26-0-163-134' +Master node: ip-26-0-163-134 ++ echo 'All nodes: ip-26-0-163-134 +ip-26-0-163-158 +ip-26-0-164-0 +ip-26-0-165-164 +ip-26-0-167-111 +ip-26-0-167-175 +ip-26-0-167-217 +ip-26-0-168-30' +All nodes: ip-26-0-163-134 +ip-26-0-163-158 +ip-26-0-164-0 +ip-26-0-165-164 +ip-26-0-167-111 +ip-26-0-167-175 +ip-26-0-167-217 +ip-26-0-168-30 ++ echo 'World size: 64' +World size: 64 ++ '[' -n '' ']' ++ srun --wait=0 --kill-on-bad-exit=1 torchrun --nnodes=8 --nproc_per_node=8 --rdzv_id=14471696 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-163-134:12356 --max_restarts 0 --rdzv_conf timeout=60 /fsx/nouamane/projects/nanotron/run_train.py --config-file benchmark/configs/config_80G_dp1_tp4_pp16_acc256_mbs1_seq4096_zero0_tpmodeRED_vocab131k.yaml +[2025-01-23 11:11:24,914] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-23 11:11:24,917] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-23 11:11:24,920] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-23 11:11:24,915] torch.distributed.run: [WARNING] +[2025-01-23 11:11:24,915] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:24,915] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-23 11:11:24,915] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:24,959] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-23 11:11:24,920] torch.distributed.run: [WARNING] +[2025-01-23 11:11:24,920] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:24,920] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-23 11:11:24,920] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:24,917] torch.distributed.run: [WARNING] +[2025-01-23 11:11:24,917] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:24,917] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-23 11:11:24,917] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:24,959] torch.distributed.run: [WARNING] +[2025-01-23 11:11:24,959] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:24,959] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-23 11:11:24,959] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:25,012] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-23 11:11:25,012] torch.distributed.run: [WARNING] +[2025-01-23 11:11:25,012] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:25,012] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-23 11:11:25,012] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:25,162] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-23 11:11:25,185] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-23 11:11:25,162] torch.distributed.run: [WARNING] +[2025-01-23 11:11:25,162] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:25,162] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-23 11:11:25,162] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:25,185] torch.distributed.run: [WARNING] +[2025-01-23 11:11:25,185] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:25,185] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-23 11:11:25,185] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:29,238] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2025-01-23 11:11:29,238] torch.distributed.run: [WARNING] +[2025-01-23 11:11:29,238] torch.distributed.run: [WARNING] ***************************************** +[2025-01-23 11:11:29,238] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2025-01-23 11:11:29,238] torch.distributed.run: [WARNING] ***************************************** +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Config: +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Config(general=GeneralArgs(project='debug', +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: run='80G_dp1_tp4_pp16_acc256_mbs1_seq4096_zero0_tpmodeRED_vocab131k', +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: seed=42, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: step=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: consumed_train_samples=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: benchmark_csv_path=PosixPath('benchmark/results/bench_final2.csv'), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: ignore_sanity_checks=True), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: parallelism=ParallelismArgs(dp=1, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pp=16, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tp=4, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pp_engine=, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tp_mode=, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tp_linear_async_communication=True, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: recompute_layer=False, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tp_recompute_allgather=True, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: expert_parallel_size=1), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: eos_token_id=0, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: hidden_act='silu', +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: hidden_size=8192, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: initializer_range=0.02, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: intermediate_size=28672, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: is_llama_config=True, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: max_position_embeddings=4096, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_attention_heads=64, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_hidden_layers=80, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_key_value_heads=64, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pad_token_id=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pretraining_tp=1, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rms_norm_eps=1e-05, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_scaling=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_theta=10000.0, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_interleaved=False, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tie_word_embeddings=False, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: use_cache=True, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: vocab_size=131072), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: init_method=RandomInit(std=0.02), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: dtype=torch.bfloat16, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: make_vocab_size_divisible_by=1, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: ddp_bucket_cap_mb=25), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tokenizer_revision=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tokenizer_max_length=None), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: checkpoint_interval=10000, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: save_initial_state=False, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: save_final_state=False, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: resume_checkpoint_path=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: checkpoints_path_is_shared_file_system=False), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: logging=LoggingArgs(log_level='info', +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: log_level_replica='info', +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: iteration_step_info_interval=1), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tokens=TokensArgs(sequence_length=4096, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: train_steps=100, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: micro_batch_size=1, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: batch_accumulation_per_replica=256, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: val_check_interval=100, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: limit_val_batches=0, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: limit_test_batches=0), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: adam_beta1=0.9, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: adam_beta2=0.95, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: torch_adam_is_fused=True, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: name='adamW'), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: zero_stage=0, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: weight_decay=0.01, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: clip_grad=1.0, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: accumulate_grad_in_fp32=True, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lr_warmup_steps=2, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lr_warmup_style='linear', +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lr_decay_style='cosine', +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lr_decay_steps=13, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lr_decay_starting_step=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: min_decay_lr=1e-05)), +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: start_training_step=1, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: data=DataArgs(dataset=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: seed=42, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_loading_workers=1))], +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: profiler=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: lighteval=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: s3_upload=None) +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Model Config: +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: LlamaConfig(bos_token_id=0, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: eos_token_id=0, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: hidden_act='silu', +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: hidden_size=8192, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: initializer_range=0.02, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: intermediate_size=28672, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: is_llama_config=True, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: max_position_embeddings=4096, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_attention_heads=64, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_hidden_layers=80, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: num_key_value_heads=64, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pad_token_id=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: pretraining_tp=1, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rms_norm_eps=1e-05, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_scaling=None, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_theta=10000.0, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: rope_interleaved=False, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: tie_word_embeddings=False, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: use_cache=True, +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: vocab_size=131072) +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Building model.. +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Initialize RoPE Theta = 10000.0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Setting PP block ranks... +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.token_position_embeddings | PP: 0/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.0 | PP: 0/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.1 | PP: 0/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.2 | PP: 0/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.3 | PP: 0/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.4 | PP: 0/16 | Block rank: 5 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.5 | PP: 0/16 | Block rank: 6 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.6 | PP: 1/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.7 | PP: 1/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.8 | PP: 1/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.9 | PP: 1/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.10 | PP: 1/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.11 | PP: 2/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.12 | PP: 2/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.13 | PP: 2/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.14 | PP: 2/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.15 | PP: 2/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.16 | PP: 3/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.17 | PP: 3/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.18 | PP: 3/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.19 | PP: 3/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.20 | PP: 3/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.21 | PP: 4/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.22 | PP: 4/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.23 | PP: 4/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.24 | PP: 4/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.25 | PP: 4/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.26 | PP: 5/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.27 | PP: 5/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.28 | PP: 5/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.29 | PP: 5/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.30 | PP: 5/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.31 | PP: 6/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.32 | PP: 6/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.33 | PP: 6/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.34 | PP: 6/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.35 | PP: 6/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.36 | PP: 7/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.37 | PP: 7/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.38 | PP: 7/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.39 | PP: 7/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.40 | PP: 7/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.41 | PP: 8/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.42 | PP: 8/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.43 | PP: 8/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.44 | PP: 8/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.45 | PP: 8/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.46 | PP: 9/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.47 | PP: 9/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.48 | PP: 9/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.49 | PP: 9/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.50 | PP: 9/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.51 | PP: 10/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.52 | PP: 10/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.53 | PP: 10/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.54 | PP: 10/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.55 | PP: 10/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.56 | PP: 11/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.57 | PP: 11/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.58 | PP: 11/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.59 | PP: 11/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.60 | PP: 11/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.61 | PP: 12/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.62 | PP: 12/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.63 | PP: 12/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.64 | PP: 12/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.65 | PP: 12/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.66 | PP: 13/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.67 | PP: 13/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.68 | PP: 13/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.69 | PP: 13/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.70 | PP: 13/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.71 | PP: 14/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.72 | PP: 14/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.73 | PP: 14/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.74 | PP: 14/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.75 | PP: 14/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.76 | PP: 14/16 | Block rank: 5 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.77 | PP: 15/16 | Block rank: 0 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.78 | PP: 15/16 | Block rank: 1 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.decoder.79 | PP: 15/16 | Block rank: 2 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.final_layer_norm | PP: 15/16 | Block rank: 3 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.lm_head | PP: 15/16 | Block rank: 4 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: model.cast_to_fp32 | PP: 15/16 | Block rank: 5 +01/23/2025 11:12:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: module_name: loss | PP: 15/16 | Block rank: 6 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=1|TP=0|ip-26-0-163-134]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Total number of parameters: 80G (152586.06MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Local number of parameters: 1.73G (3296.19MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=9|TP=0|ip-26-0-167-111]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=8|TP=0|ip-26-0-167-111]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=13|TP=0|ip-26-0-167-217]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=12|TP=0|ip-26-0-167-217]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=14|TP=0|ip-26-0-168-30]: Local number of parameters: 1.46G (2784.19MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=15|TP=0|ip-26-0-168-30]: Local number of parameters: 998M (1904.11MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=1|TP=0|ip-26-0-163-134]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=9|TP=0|ip-26-0-167-111]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [After model building] Memory usage: 3296.20MiB. Peak allocated: 3296.23MiB Peak reserved: 3298.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=8|TP=0|ip-26-0-167-111]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=13|TP=0|ip-26-0-167-217]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=12|TP=0|ip-26-0-167-217]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=15|TP=0|ip-26-0-168-30]: [After model building] Memory usage: 1904.12MiB. Peak allocated: 1904.15MiB Peak reserved: 1906.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=14|TP=0|ip-26-0-168-30]: [After model building] Memory usage: 2784.20MiB. Peak allocated: 2784.23MiB Peak reserved: 2786.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: No checkpoint path provided. +01/23/2025 11:12:25 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Parametrizing model parameters using StandardParametrizator +01/23/2025 11:12:25 [INFO|DP=0|PP=11|TP=0|ip-26-0-167-175]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=10|TP=0|ip-26-0-167-175]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=11|TP=0|ip-26-0-167-175]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=10|TP=0|ip-26-0-167-175]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=5|TP=0|ip-26-0-164-0]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=4|TP=0|ip-26-0-164-0]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=5|TP=0|ip-26-0-164-0]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=4|TP=0|ip-26-0-164-0]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=7|TP=0|ip-26-0-165-164]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=1|TP=2|ip-26-0-163-134]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=0|TP=2|ip-26-0-163-134]: Local number of parameters: 1.73G (3296.19MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=6|TP=0|ip-26-0-165-164]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=7|TP=0|ip-26-0-165-164]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=1|TP=2|ip-26-0-163-134]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=0|TP=2|ip-26-0-163-134]: [After model building] Memory usage: 3296.20MiB. Peak allocated: 3296.23MiB Peak reserved: 3298.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=0|TP=1|ip-26-0-163-134]: Local number of parameters: 1.73G (3296.19MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=6|TP=0|ip-26-0-165-164]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=1|TP=1|ip-26-0-163-134]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=3|TP=0|ip-26-0-163-158]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=2|TP=0|ip-26-0-163-158]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=0|TP=3|ip-26-0-163-134]: Local number of parameters: 1.73G (3296.19MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=0|TP=1|ip-26-0-163-134]: [After model building] Memory usage: 3296.20MiB. Peak allocated: 3296.23MiB Peak reserved: 3298.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=1|TP=1|ip-26-0-163-134]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=3|TP=0|ip-26-0-163-158]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=1|TP=3|ip-26-0-163-134]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=2|TP=0|ip-26-0-163-158]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=0|TP=3|ip-26-0-163-134]: [After model building] Memory usage: 3296.20MiB. Peak allocated: 3296.23MiB Peak reserved: 3298.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=1|TP=3|ip-26-0-163-134]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=5|TP=2|ip-26-0-164-0]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=4|TP=2|ip-26-0-164-0]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=9|TP=2|ip-26-0-167-111]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=2|TP=2|ip-26-0-163-158]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=3|TP=2|ip-26-0-163-158]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=5|TP=2|ip-26-0-164-0]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=8|TP=2|ip-26-0-167-111]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=4|TP=2|ip-26-0-164-0]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=9|TP=2|ip-26-0-167-111]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=2|TP=2|ip-26-0-163-158]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=3|TP=2|ip-26-0-163-158]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=2|TP=1|ip-26-0-163-158]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=7|TP=2|ip-26-0-165-164]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=8|TP=1|ip-26-0-167-111]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=8|TP=2|ip-26-0-167-111]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=3|TP=1|ip-26-0-163-158]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=9|TP=1|ip-26-0-167-111]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=4|TP=1|ip-26-0-164-0]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=5|TP=1|ip-26-0-164-0]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=7|TP=2|ip-26-0-165-164]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=2|TP=1|ip-26-0-163-158]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=3|TP=3|ip-26-0-163-158]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=5|TP=3|ip-26-0-164-0]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=8|TP=1|ip-26-0-167-111]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=6|TP=2|ip-26-0-165-164]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=8|TP=3|ip-26-0-167-111]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=4|TP=3|ip-26-0-164-0]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=9|TP=1|ip-26-0-167-111]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=3|TP=1|ip-26-0-163-158]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=4|TP=1|ip-26-0-164-0]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=5|TP=1|ip-26-0-164-0]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=9|TP=3|ip-26-0-167-111]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=3|TP=3|ip-26-0-163-158]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=5|TP=3|ip-26-0-164-0]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=8|TP=3|ip-26-0-167-111]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=7|TP=1|ip-26-0-165-164]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=6|TP=1|ip-26-0-165-164]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=4|TP=3|ip-26-0-164-0]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=2|TP=3|ip-26-0-163-158]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=6|TP=2|ip-26-0-165-164]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=9|TP=3|ip-26-0-167-111]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=7|TP=3|ip-26-0-165-164]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=7|TP=1|ip-26-0-165-164]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=6|TP=1|ip-26-0-165-164]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=6|TP=3|ip-26-0-165-164]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=2|TP=3|ip-26-0-163-158]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=7|TP=3|ip-26-0-165-164]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=13|TP=2|ip-26-0-167-217]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=6|TP=3|ip-26-0-165-164]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=11|TP=2|ip-26-0-167-175]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=12|TP=2|ip-26-0-167-217]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=10|TP=2|ip-26-0-167-175]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=13|TP=2|ip-26-0-167-217]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=11|TP=2|ip-26-0-167-175]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=12|TP=2|ip-26-0-167-217]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=10|TP=2|ip-26-0-167-175]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=11|TP=1|ip-26-0-167-175]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=12|TP=1|ip-26-0-167-217]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=10|TP=1|ip-26-0-167-175]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=14|TP=2|ip-26-0-168-30]: Local number of parameters: 1.46G (2784.19MiB) +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=13|TP=1|ip-26-0-167-217]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=15|TP=2|ip-26-0-168-30]: Local number of parameters: 998M (1904.11MiB) +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=11|TP=1|ip-26-0-167-175]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=13|TP=3|ip-26-0-167-217]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=10|TP=3|ip-26-0-167-175]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=12|TP=3|ip-26-0-167-217]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=11|TP=3|ip-26-0-167-175]: Local number of parameters: 1.22G (2320.16MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=10|TP=1|ip-26-0-167-175]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=12|TP=1|ip-26-0-167-217]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=14|TP=2|ip-26-0-168-30]: [After model building] Memory usage: 2784.20MiB. Peak allocated: 2784.23MiB Peak reserved: 2786.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=13|TP=1|ip-26-0-167-217]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=15|TP=2|ip-26-0-168-30]: [After model building] Memory usage: 1904.12MiB. Peak allocated: 1904.15MiB Peak reserved: 1906.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=15|TP=1|ip-26-0-168-30]: Local number of parameters: 998M (1904.11MiB) +01/23/2025 11:12:25 [INFO|DP=0|PP=13|TP=3|ip-26-0-167-217]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=10|TP=3|ip-26-0-167-175]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=12|TP=3|ip-26-0-167-217]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +01/23/2025 11:12:25 [INFO|DP=0|PP=11|TP=3|ip-26-0-167-175]: [After model building] Memory usage: 2320.17MiB. Peak allocated: 2320.19MiB Peak reserved: 2322.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=15|TP=1|ip-26-0-168-30]: [After model building] Memory usage: 1904.12MiB. Peak allocated: 1904.15MiB Peak reserved: 1906.00MiB +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=14|TP=3|ip-26-0-168-30]: Local number of parameters: 1.46G (2784.19MiB) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=14|TP=1|ip-26-0-168-30]: Local number of parameters: 1.46G (2784.19MiB) +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=15|TP=3|ip-26-0-168-30]: Local number of parameters: 998M (1904.11MiB) +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=14|TP=3|ip-26-0-168-30]: [After model building] Memory usage: 2784.20MiB. Peak allocated: 2784.23MiB Peak reserved: 2786.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=15|TP=3|ip-26-0-168-30]: [After model building] Memory usage: 1904.12MiB. Peak allocated: 1904.15MiB Peak reserved: 1906.00MiB +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:25 [INFO|DP=0|PP=14|TP=1|ip-26-0-168-30]: [After model building] Memory usage: 2784.20MiB. Peak allocated: 2784.23MiB Peak reserved: 2786.00MiB +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +NCCL version 2.18.5+cuda12.2 +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [Optimizer Building] Using LearningRateForSP as learning rate +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Before merge_named_param_groups +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Memory usage: 3296.20MiB. Peak allocated 3296.23MiB. Peak reserved: 3298.00MiB +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: After merge_named_param_groups +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Memory usage: 3296.20MiB. Peak allocated 3296.20MiB. Peak reserved: 3298.00MiB +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Before fp32_grad_optimizer_builder +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Memory usage: 3296.20MiB. Peak allocated 3296.20MiB. Peak reserved: 3298.00MiB +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Before basic_optimizer_builder +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Memory usage: 16480.95MiB. Peak allocated 16480.95MiB. Peak reserved: 16486.00MiB +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: After basic_optimizer_builder +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Memory usage: 16480.95MiB. Peak allocated 16480.95MiB. Peak reserved: 16486.00MiB +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: After fp32_grad_optimizer_builder +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Memory usage: 16480.95MiB. Peak allocated 16480.95MiB. Peak reserved: 16486.00MiB +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Using dummy data generator +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [Training Plan] There are 1 training stages +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [Stage Stable Training Stage] start from step 1 +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: +01/23/2025 11:12:26 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: [Start training] datetime: 2025-01-23 11:12:26.985760 | mbs: 1 | grad_accum: 256 | global_batch_size: 256 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +01/23/2025 11:12:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +01/23/2025 11:12:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Before TBI +01/23/2025 11:12:27 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Memory usage: 16480.95MiB. Peak allocated 16480.95MiB. Peak reserved: 16486.00MiB +01/23/2025 11:12:27 [WARNING|DP=0|PP=0|TP=0|ip-26-0-163-134]: We torch.cuda.empty_cache() before every forward pass to avoid OOM +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +[W ProcessGroupNCCL.cpp:1856] Warning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (function operator()) +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: 0NCCL_AVOID_RECORD_STREAMS=1 has no effect for point-to-point collectives. (Triggered internally at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1856.) + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass +01/23/2025 11:14:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: After TBI +01/23/2025 11:14:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Memory usage: 16678.99MiB. Peak allocated 54779.96MiB. Peak reserved: 57426.00MiB +01/23/2025 11:14:28 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Memory usage: 16678.99MiB. Peak allocated 16678.99MiB. Peak reserved: 57394.00MiB +01/23/2025 11:14:38 [INFO|DP=0|PP=15|TP=0|ip-26-0-168-30]: iteration: 1 / 100 | consumed_tokens: 1.05M | elapsed_time_per_iteration_ms: 130K | tokens_per_sec: 8.08K | tokens_per_sec_per_gpu: 126 | global_batch_size: 256 | lm_loss: 13.4 | lr: 0.00015 | model_tflops_per_gpu: 63.9 | hardware_tflops_per_gpu: 63.9 | grad_norm: 13.6 | cuda_memory_allocated: 18.2G | cuda_max_memory_reserved: 20.1G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.5G | hd_free_memory_tb: 243G +01/23/2025 11:14:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Before TBI +01/23/2025 11:14:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-163-134]: Memory usage: 29863.76MiB. Peak allocated 29863.76MiB. Peak reserved: 68146.00MiB +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 254, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 464, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 505, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 345, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 104, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 342, in backward + grad_weight = grad_output.t().matmul(total_tensor) +torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 224.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 189.94 MiB is free. Including non-PyTorch memory, this process has 79.13 GiB memory in use. Of the allocated memory 65.53 GiB is allocated by PyTorch, and 965.56 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF +[E ProcessGroupGloo.cpp:2810] [Rank 0]: Rank 1 failed to pass monitoredBarrier in 1200000 ms +[E ProcessGroupGloo.cpp:138] [Rank 0]: Ranks 1 failed to pass monitoredBarrier in 1200000 ms +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 254, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 464, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 505, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 345, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 104, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 508, in backward + handle = dist.all_gather_into_tensor(total_grad_output, grad_output, group=group, async_op=True) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 60, in all_gather_into_tensor + return dist.all_gather_into_tensor( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2897, in all_gather_into_tensor + work = group._allgather_base(output_tensor, input_tensor) +RuntimeError: [Rank 0]: Ranks 1 failed to pass monitoredBarrier in 1200000 ms +[E ProcessGroupGloo.cpp:138] Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +[E ProcessGroupGloo.cpp:138] Rank 3 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. +[E TraceUtils.h:35] Store is down while updating #9691 with key NCCL_0_trace_end +[E TraceUtils.h:35] Store is down while updating #13418 with key NCCL_2_trace_end +[E TraceUtils.h:35] Store is down while updating #9698 with key NCCL_3_trace_start +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 254, in +Traceback (most recent call last): + File "/fsx/nouamane/projects/nanotron/run_train.py", line 254, in + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + trainer.train(dataloader) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 464, in train + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 464, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 505, in training_step + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 505, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 345, in train_batch_iter + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 345, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 104, in backward + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 104, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 508, in backward + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 508, in backward + handle = dist.all_gather_into_tensor(total_grad_output, grad_output, group=group, async_op=True) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 60, in all_gather_into_tensor + handle = dist.all_gather_into_tensor(total_grad_output, grad_output, group=group, async_op=True) + File "/fsx/nouamane/projects/nanotron/src/nanotron/distributed.py", line 60, in all_gather_into_tensor + return dist.all_gather_into_tensor( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return dist.all_gather_into_tensor( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2897, in all_gather_into_tensor + return func(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 2897, in all_gather_into_tensor + work = group._allgather_base(output_tensor, input_tensor) +RuntimeError: Rank 3 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.163.134]:2450 + work = group._allgather_base(output_tensor, input_tensor) +RuntimeError: Rank 2 successfully reached monitoredBarrier, but received errors while waiting for send/recv from rank 0. Please check rank 0 logs for faulty rank. + Original exception: +[../third_party/gloo/gloo/transport/tcp/pair.cc:534] Connection closed by peer [26.0.163.134]:28307 +[E TraceUtils.h:35] Store is down while updating #9698 with key NCCL_1_trace_start +[E TraceUtils.h:35] Store is down while updating #13446 with key NCCL_0_trace_end +[E TraceUtils.h:35] Store is down while updating #13467 with key NCCL_0_trace_end +[E TraceUtils.h:35] Store is down while updating #13467 with key NCCL_3_trace_end +[E TraceUtils.h:35] Store is down while updating #13467 with key NCCL_2_trace_end +[E TraceUtils.h:35] Store is down while updating #13477 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #13477 with key NCCL_3_trace_start +[E TraceUtils.h:35] Store is down while updating #16195 with key NCCL_0_trace_start +[E TraceUtils.h:35] Store is down while updating #16243 with key NCCL_2_trace_end +[E TraceUtils.h:35] Store is down while updating #13552 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #16259 with key NCCL_1_trace_end +[E TraceUtils.h:35] Store is down while updating #16264 with key NCCL_3_trace_start +[2025-01-23 11:14:55,532] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 182391 closing signal SIGTERM +[2025-01-23 11:14:55,533] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 182393 closing signal SIGTERM +[2025-01-23 11:14:55,533] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 182394 closing signal SIGTERM +[2025-01-23 11:14:55,533] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 182395 closing signal SIGTERM +[2025-01-23 11:14:55,533] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 182396 closing signal SIGTERM +[2025-01-23 11:14:55,533] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 182397 closing signal SIGTERM +[2025-01-23 11:14:55,533] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 182398 closing signal SIGTERM +[2025-01-23 11:15:15,335] torch.distributed.elastic.multiprocessing.api: [ERROR] failed (exitcode: 1) local_rank: 1 (pid: 182392) of binary: /fsx/nouamane/miniconda/envs/2-1-cu121/bin/python +[2025-01-23 11:15:15,369] torch.distributed.elastic.multiprocessing.errors.error_handler: [ERROR] no error file defined for parent, to copy child error file (/tmp/torchelastic_mwsir5oz/14471696_o8sbktc2/attempt_0/1/error.json) +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 264, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +/fsx/nouamane/projects/nanotron/run_train.py FAILED +------------------------------------------------------------ +Failures: + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2025-01-23_11:14:41 + host : ip-26-0-163-134.ec2.internal + rank : 1 (local_rank: 1) + exitcode : 1 (pid: 182392) + error_file: /tmp/torchelastic_mwsir5oz/14471696_o8sbktc2/attempt_0/1/error.json + traceback : Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 464, in train + outputs, loss_avg = self.training_step(dataloader=self.current_dataloader) + File "/fsx/nouamane/projects/nanotron/src/nanotron/trainer.py", line 505, in training_step + outputs = self.pipeline_engine.train_batch_iter( + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 345, in train_batch_iter + self.backward(context=context, state=state, grad_accumulator=grad_accumulator) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/pipeline_parallel/engine.py", line 104, in backward + grad_accumulator.backward(activations) + File "/fsx/nouamane/projects/nanotron/src/nanotron/optim/gradient_accumulator.py", line 207, in backward + loss.backward() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/_tensor.py", line 492, in backward + torch.autograd.backward( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/__init__.py", line 251, in backward + Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/autograd/function.py", line 288, in apply + return user_fn(self, *args) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/utils.py", line 40, in wrapper + return func(*args, **kwargs) + File "/fsx/nouamane/projects/nanotron/src/nanotron/parallel/tensor_parallel/functional.py", line 342, in backward + grad_weight = grad_output.t().matmul(total_tensor) + torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 224.00 MiB. GPU 1 has a total capacty of 79.33 GiB of which 189.94 MiB is free. Including non-PyTorch memory, this process has 79.13 GiB memory in use. Of the allocated memory 65.53 GiB is allocated by PyTorch, and 965.56 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF + +============================================================ +[2025-01-23 11:15:15,484] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-165-164.ec2.internal_32790_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:15,554] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32868 closing signal SIGTERM +[2025-01-23 11:15:15,554] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32869 closing signal SIGTERM +[2025-01-23 11:15:15,554] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32870 closing signal SIGTERM +[2025-01-23 11:15:15,554] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32871 closing signal SIGTERM +[2025-01-23 11:15:15,554] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32872 closing signal SIGTERM +[2025-01-23 11:15:15,554] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32873 closing signal SIGTERM +[2025-01-23 11:15:15,554] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32874 closing signal SIGTERM +[2025-01-23 11:15:15,556] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 32875 closing signal SIGTERM +srun: error: ip-26-0-163-134: task 0: Exited with exit code 1 +srun: launch/slurm: _step_signal: Terminating StepId=14471696.0 +[2025-01-23 11:15:16,215] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-23 11:15:16,215] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3511952 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2423015 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3511953 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2423016 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-23 11:15:16,215] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2935159 closing signal SIGTERM +[2025-01-23 11:15:16,215] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2423017 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2754171 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 576580 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2935160 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1353362 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 576581 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2423018 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2754172 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2754173 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2935161 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 576582 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1353363 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1353364 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 576583 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2935162 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2935163 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1353365 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2423019 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2935164 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 576584 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 576585 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 576586 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 576587 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3511954 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3511955 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3511956 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3511957 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3511958 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1353366 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1353367 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1353368 closing signal SIGTERM +[2025-01-23 11:15:16,216] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1353369 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2754174 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2754175 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2754176 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2754177 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2754178 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2935165 closing signal SIGTERM +[2025-01-23 11:15:16,217] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3511959 closing signal SIGTERM +[2025-01-23 11:15:16,220] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2423020 closing signal SIGTERM +[2025-01-23 11:15:16,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2423021 closing signal SIGTERM +[2025-01-23 11:15:16,222] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2423022 closing signal SIGTERM +[2025-01-23 11:15:16,221] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2935166 closing signal SIGTERM +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/rendezvous/c10d_rendezvous_backend.py", line 113, in _call_store + return getattr(self._store, store_op)(*args, **kwargs) +RuntimeError: Broken pipe + +The above exception was the direct cause of the following exception: + +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 909, in _invoke_run + num_nodes_waiting = rdzv_handler.num_nodes_waiting() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/rendezvous/dynamic_rendezvous.py", line 1083, in num_nodes_waiting + self._state_holder.sync() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/rendezvous/dynamic_rendezvous.py", line 409, in sync + get_response = self._backend.get_state() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/rendezvous/c10d_rendezvous_backend.py", line 73, in get_state + base64_state: bytes = self._call_store("get", self._key) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/rendezvous/c10d_rendezvous_backend.py", line 115, in _call_store + raise RendezvousConnectionError( +torch.distributed.elastic.rendezvous.api.RendezvousConnectionError: The connection to the C10d store has failed. See inner exception for details. + +During handling of the above exception, another exception occurred: + +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 748, in run + self._shutdown() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/local_elastic_agent.py", line 289, in _shutdown + self._pcontext.close(death_sig) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 331, in close + self._close(death_sig=death_sig, timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 713, in _close + handler.proc.wait(time_to_wait) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1209, in wait + return self._wait(timeout=timeout) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/subprocess.py", line 1953, in _wait + time.sleep(delay) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 32790 got signal: 15 +[2025-01-23 11:15:19,636] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-111.ec2.internal_2754093_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:20,126] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-175.ec2.internal_576503_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:20,206] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-0.ec2.internal_1353284_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:20,276] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-30.ec2.internal_2422936_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:20,296] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-217.ec2.internal_2935082_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:20,391] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-163-158.ec2.internal_3511874_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:24,638] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-111.ec2.internal_2754093_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:25,127] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-175.ec2.internal_576503_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:25,208] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-0.ec2.internal_1353284_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:25,278] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-30.ec2.internal_2422936_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:25,298] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-217.ec2.internal_2935082_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:25,393] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-163-158.ec2.internal_3511874_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:29,640] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-111.ec2.internal_2754093_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:30,129] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-175.ec2.internal_576503_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:30,209] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-0.ec2.internal_1353284_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:30,279] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-30.ec2.internal_2422936_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:30,299] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-217.ec2.internal_2935082_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:30,394] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-163-158.ec2.internal_3511874_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:34,641] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-111.ec2.internal_2754093_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:35,130] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-175.ec2.internal_576503_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:35,210] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-0.ec2.internal_1353284_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:35,281] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-168-30.ec2.internal_2422936_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:35,301] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-217.ec2.internal_2935082_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:35,396] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-163-158.ec2.internal_3511874_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2422936 got signal: 15 +srun: error: ip-26-0-168-30: task 7: Exited with exit code 1 +[2025-01-23 11:15:39,642] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-111.ec2.internal_2754093_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:40,132] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-175.ec2.internal_576503_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:40,212] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-0.ec2.internal_1353284_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:40,302] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-217.ec2.internal_2935082_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:40,397] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-163-158.ec2.internal_3511874_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +[2025-01-23 11:15:44,644] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-167-111.ec2.internal_2754093_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2935082 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 576503 got signal: 15 +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 2754093 got signal: 15 +[2025-01-23 11:15:45,213] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-164-0.ec2.internal_1353284_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +srun: error: ip-26-0-167-217: task 6: Exited with exit code 1 +srun: error: ip-26-0-167-175: task 5: Exited with exit code 1 +[2025-01-23 11:15:45,398] torch.distributed.elastic.rendezvous.dynamic_rendezvous: [WARNING] The node 'ip-26-0-163-158.ec2.internal_3511874_0' has failed to send a keep-alive heartbeat to the rendezvous '14471696' due to an error of type RendezvousConnectionError. +Traceback (most recent call last): + File "/fsx/nouamane/miniconda/envs/2-1-cu121/bin/torchrun", line 8, in + sys.exit(main()) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 346, in wrapper + return f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 806, in main + run(args) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/run.py", line 797, in run + elastic_launch( + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 134, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/launcher/api.py", line 255, in launch_agent + result = agent.run() + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/metrics/api.py", line 124, in wrapper + result = f(*args, **kwargs) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 736, in run + result = self._invoke_run(role) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/agent/server/api.py", line 877, in _invoke_run + time.sleep(monitor_interval) + File "/fsx/nouamane/miniconda/envs/2-1-cu121/lib/python3.10/site-packages/torch/distributed/elastic/multiprocessing/api.py", line 62, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 1353284 got signal: 15 +srun: error: ip-26-0-167-111: task 4: Exited with exit code 1 +srun: error: ip-26-0-164-0: task 2: Exited with exit code 1 +srun: error: ip-26-0-163-158: task 1: Killed +srun: error: ip-26-0-165-164: task 3: Exited with exit code 1 +srun: Force Terminated StepId=14471696.0 diff --git a/logs/bench_dp128_tp1_pp1_acc4_mbs2_seq4096_zero0_l24_h2048_heads16-12178054.out b/logs/bench_dp128_tp1_pp1_acc4_mbs2_seq4096_zero0_l24_h2048_heads16-12178054.out new file mode 100644 index 0000000000000000000000000000000000000000..b414163d219ccea3f6dafac27abb94bd9002dccb --- /dev/null +++ b/logs/bench_dp128_tp1_pp1_acc4_mbs2_seq4096_zero0_l24_h2048_heads16-12178054.out @@ -0,0 +1,975 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/admin/home/nouamane/.local/share/man:/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/opt/amazon/openmpi/share/man:/opt/amazon/efa/share/man:/opt/slurm/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/cuda-12.1/include:1:/opt/amazon/efa/bin:1:/admin/home/nouamane/.local/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/usr/sbin:1:/fsx/nouamane/miniconda/bin:1:/opt/slurm/bin:1:/usr/games:1:/usr/local/sbin:1:/usr/local/cuda-12.1/bin:2:/opt/amazon/openmpi/bin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' +++ eval '. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh" +PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda'\'' +export CONDA_SHLVL='\''3'\'' +export CONDA_DEFAULT_ENV='\''base'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\'' +export CONDA_PREFIX_2='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\''' ++++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/deactivate.d/libxml2_deactivate.sh +++++ test -n '' +++++ unset XML_CATALOG_FILES +++++ unset xml_catalog_files_libxml2 ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++++ export CONDA_PREFIX=/fsx/nouamane/miniconda ++++ CONDA_PREFIX=/fsx/nouamane/miniconda ++++ export CONDA_SHLVL=3 ++++ CONDA_SHLVL=3 ++++ export CONDA_DEFAULT_ENV=base ++++ CONDA_DEFAULT_ENV=base ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' ++++ export CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ CONDA_PREFIX_2=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''4'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_3='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=4 +++ CONDA_SHLVL=4 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ CONDA_PREFIX_3=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/admin/home/nouamane/.local/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/opt/slurm/bin:/usr/local/cuda-12.1/bin:/usr/local/cuda-12.1/include:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/amazon/efa/bin:/opt/amazon/openmpi/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin +++ scontrol show hostnames 'ip-26-0-160-[100,225,242],ip-26-0-161-78,ip-26-0-163-[147,226],ip-26-0-170-[31,132,143,160],ip-26-0-173-[202,246],ip-26-0-174-[196,240],ip-26-0-175-[19,34]' ++ export 'NODELIST=ip-26-0-160-100 +ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-163-147 +ip-26-0-163-226 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34' ++ NODELIST='ip-26-0-160-100 +ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-163-147 +ip-26-0-163-226 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-160-[100,225,242],ip-26-0-161-78,ip-26-0-163-[147,226],ip-26-0-170-[31,132,143,160],ip-26-0-173-[202,246],ip-26-0-174-[196,240],ip-26-0-175-[19,34]' ++ export MASTER_NODE=ip-26-0-160-100 ++ MASTER_NODE=ip-26-0-160-100 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=16 ++ NNODES=16 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=128 ++ WORLD_SIZE=128 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-100' +Master node: ip-26-0-160-100 ++ echo 'All nodes: ip-26-0-160-100 +ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-163-147 +ip-26-0-163-226 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34' +All nodes: ip-26-0-160-100 +ip-26-0-160-225 +ip-26-0-160-242 +ip-26-0-161-78 +ip-26-0-163-147 +ip-26-0-163-226 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34 ++ echo 'World size: 128' +World size: 128 ++ srun torchrun --nnodes=16 --nproc_per_node=8 --rdzv_id=12178054 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-100:12356 run_train.py --config-file benchmark/configs/config_dp128_tp1_pp1_acc4_mbs2_seq4096_zero0_l24_h2048_heads16.yaml +[2024-12-02 10:25:48,119] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:48,124] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:48,119] torch.distributed.run: [WARNING] +[2024-12-02 10:25:48,119] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:48,119] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:48,119] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:48,124] torch.distributed.run: [WARNING] +[2024-12-02 10:25:48,124] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:48,124] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:48,124] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,508] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,507] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,659] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,666] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,698] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,659] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,659] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,659] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,659] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,666] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,666] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,666] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,666] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,698] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,698] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,698] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,698] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,761] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,779] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,761] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,761] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,761] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,761] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,815] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:52,815] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,815] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,815] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,815] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,779] torch.distributed.run: [WARNING] +[2024-12-02 10:25:52,779] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:52,779] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:52,779] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:53,026] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-12-02 10:25:53,026] torch.distributed.run: [WARNING] +[2024-12-02 10:25:53,026] torch.distributed.run: [WARNING] ***************************************** +[2024-12-02 10:25:53,026] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-12-02 10:25:53,026] torch.distributed.run: [WARNING] ***************************************** +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Config: +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Config(general=GeneralArgs(project='debug', +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: run='dp128_tp1_pp1_acc4_mbs2_seq4096_zero0_l24_h2048_heads16', +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: seed=42, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: step=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: consumed_train_samples=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: benchmark_csv_path=PosixPath('bench.csv'), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: ignore_sanity_checks=True), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: parallelism=ParallelismArgs(dp=128, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pp=1, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp=1, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pp_engine=, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp_mode=, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp_linear_async_communication=True, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: recompute_layer=False, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tp_recompute_allgather=True, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: expert_parallel_size=1), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: eos_token_id=0, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_act='silu', +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_size=2048, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: initializer_range=0.02, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: intermediate_size=8192, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: is_llama_config=True, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: max_position_embeddings=4096, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_attention_heads=16, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_hidden_layers=24, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_key_value_heads=16, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pad_token_id=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pretraining_tp=1, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rms_norm_eps=1e-05, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_scaling=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_theta=10000.0, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_interleaved=False, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tie_word_embeddings=True, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: use_cache=True, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: vocab_size=49152), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: init_method=RandomInit(std=0.02), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: dtype=torch.bfloat16, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: make_vocab_size_divisible_by=1, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: ddp_bucket_cap_mb=25), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokenizer_revision=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokenizer_max_length=None), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: checkpoint_interval=10000, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: save_initial_state=False, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: save_final_state=False, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: resume_checkpoint_path=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: checkpoints_path_is_shared_file_system=False), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: logging=LoggingArgs(log_level='info', +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: log_level_replica='info', +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: iteration_step_info_interval=1), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tokens=TokensArgs(sequence_length=4096, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: train_steps=100, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: micro_batch_size=2, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: batch_accumulation_per_replica=4, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: val_check_interval=100, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: limit_val_batches=0, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: limit_test_batches=0), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: adam_beta1=0.9, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: adam_beta2=0.95, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: torch_adam_is_fused=True, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: name='adamW'), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: zero_stage=0, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: weight_decay=0.01, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: clip_grad=1.0, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: accumulate_grad_in_fp32=True, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_warmup_steps=2, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_warmup_style='linear', +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_decay_style='cosine', +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_decay_steps=13, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lr_decay_starting_step=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: min_decay_lr=1e-05)), +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: start_training_step=1, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: data=DataArgs(dataset=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: seed=42, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_loading_workers=1))], +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: profiler=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: lighteval=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: s3_upload=None) +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Model Config: +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: LlamaConfig(bos_token_id=0, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: eos_token_id=0, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_act='silu', +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: hidden_size=2048, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: initializer_range=0.02, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: intermediate_size=8192, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: is_llama_config=True, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: max_position_embeddings=4096, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_attention_heads=16, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_hidden_layers=24, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: num_key_value_heads=16, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pad_token_id=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: pretraining_tp=1, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rms_norm_eps=1e-05, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_scaling=None, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_theta=10000.0, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: rope_interleaved=False, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: tie_word_embeddings=True, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: use_cache=True, +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: vocab_size=49152) +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Building model.. +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Initialize RoPE Theta = 10000.0 +12/02/2024 10:26:39 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Setting PP block ranks... +12/02/2024 10:26:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Total number of parameters: 1.71G (3264.19MiB) +12/02/2024 10:26:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Local number of parameters: 1.71G (3264.19MiB) +12/02/2024 10:26:42 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: [After model building] Memory usage: 3264.22MiB. Peak allocated: 5392.00MiB Peak reserved: 29974.00MiB +12/02/2024 10:26:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: No checkpoint path provided. +12/02/2024 10:26:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Parametrizing model parameters using StandardParametrizator +12/02/2024 10:26:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: [Optimizer Building] Using LearningRateForSP as learning rate +12/02/2024 10:26:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +12/02/2024 10:26:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Using dummy data generator +12/02/2024 10:26:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: [Training Plan] There are 1 training stages +12/02/2024 10:26:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: [Stage Stable Training Stage] start from step 1 +12/02/2024 10:26:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: +12/02/2024 10:26:52 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: [Start training] datetime: 2024-12-02 10:26:52.859434 | mbs: 2 | grad_accum: 4 | global_batch_size: 1024 | sequence_length: 4096 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +12/02/2024 10:27:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +12/02/2024 10:27:01 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Memory usage: 19585.18MiB. Peak allocated 19585.18MiB. Peak reserved: 46300.00MiB +12/02/2024 10:27:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Memory usage: 19673.32MiB. Peak allocated 42791.24MiB. Peak reserved: 45144.00MiB +12/02/2024 10:27:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: iteration: 1 / 100 | consumed_tokens: 4.19M | elapsed_time_per_iteration_ms: 3.09K | tokens_per_sec: 1.36M | tokens_per_sec_per_gpu: 10.6K | global_batch_size: 1.02K | lm_loss: 11.2 | lr: 0.00015 | model_tflops_per_gpu: 134 | hardware_tflops_per_gpu: 134 | grad_norm: 0.831 | cuda_memory_allocated: 34.3G | cuda_max_memory_reserved: 47.3G | hd_total_memory_tb: 312G | hd_used_memory_tb: 71.1G | hd_free_memory_tb: 241G +12/02/2024 10:27:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Memory usage: 32730.16MiB. Peak allocated 32730.16MiB. Peak reserved: 45144.00MiB +12/02/2024 10:27:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Memory usage: 32730.16MiB. Peak allocated 55848.08MiB. Peak reserved: 58744.00MiB +12/02/2024 10:27:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: iteration: 2 / 100 | consumed_tokens: 8.39M | elapsed_time_per_iteration_ms: 1.38K | tokens_per_sec: 3.05M | tokens_per_sec_per_gpu: 23.8K | global_batch_size: 1.02K | lm_loss: 11.2 | lr: 0.0003 | model_tflops_per_gpu: 302 | hardware_tflops_per_gpu: 302 | grad_norm: 0.831 | cuda_memory_allocated: 34.3G | cuda_max_memory_reserved: 61.6G | hd_total_memory_tb: 312G | hd_used_memory_tb: 71.1G | hd_free_memory_tb: 241G +12/02/2024 10:27:06 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Memory usage: 32730.16MiB. Peak allocated 32730.24MiB. Peak reserved: 58744.00MiB +12/02/2024 10:27:07 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Memory usage: 32730.16MiB. Peak allocated 55848.08MiB. Peak reserved: 58744.00MiB +12/02/2024 10:27:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: iteration: 3 / 100 | consumed_tokens: 12.6M | elapsed_time_per_iteration_ms: 1.63K | tokens_per_sec: 2.58M | tokens_per_sec_per_gpu: 20.1K | global_batch_size: 1.02K | lm_loss: 11.2 | lr: 0.000296 | model_tflops_per_gpu: 255 | hardware_tflops_per_gpu: 255 | grad_norm: 0.799 | cuda_memory_allocated: 34.3G | cuda_max_memory_reserved: 61.6G | hd_total_memory_tb: 312G | hd_used_memory_tb: 71.1G | hd_free_memory_tb: 241G +12/02/2024 10:27:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | dp | pp | tp | pp_engine | tp_mode | tp_async_comm | hidden_size | hidden_act | num_layers | num_heads | num_kv_heads | max_pos | vocab_size | tie_word_embeddings | dtype | zero_stage | ddp_bucket_cap_mb | accumulate_grad_in_fp32 | +12/02/2024 10:27:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: | -------- | ------------------------------------------------------- | ----- | ------- | --- | ----------- | ---- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | --- | -- | -- | --------------------------------------------------------------------------------------------------------- | --------------------------------------- | ------------- | ----------- | ---------- | ---------- | --------- | ------------ | ------- | ---------- | ------------------- | -------------- | ---------- | ----------------- | ----------------------- | +12/02/2024 10:27:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: | 12178054 | dp128_tp1_pp1_acc4_mbs2_seq4096_zero0_l24_h2048_heads16 | 16 | 4096 | 2 | 4 | 1024 | 255.38 | 255.38 | 20134.72 | 192.00 | 112.31 | 115.59 | 459.55 | 265.06 | 264.99 | 31.96 | 57.37 | 128 | 1 | 1 | | TensorParallelLinearMode.REDUCE_SCATTER | True | 2048 | silu | 24 | 16 | 16 | 4096 | 49152 | True | torch.bfloat16 | 0 | 25 | True | +12/02/2024 10:27:08 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-100]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=1|PP=0|TP=0|ip-26-0-160-100]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=64|PP=0|TP=0|ip-26-0-170-160]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=65|PP=0|TP=0|ip-26-0-170-160]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=32|PP=0|TP=0|ip-26-0-163-147]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=24|PP=0|TP=0|ip-26-0-161-78]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=2|PP=0|TP=0|ip-26-0-160-100]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=4|PP=0|TP=0|ip-26-0-160-100]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=3|PP=0|TP=0|ip-26-0-160-100]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=96|PP=0|TP=0|ip-26-0-174-196]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=66|PP=0|TP=0|ip-26-0-170-160]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=68|PP=0|TP=0|ip-26-0-170-160]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=80|PP=0|TP=0|ip-26-0-173-202]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=56|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=67|PP=0|TP=0|ip-26-0-170-160]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=16|PP=0|TP=0|ip-26-0-160-242]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=49|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=48|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=25|PP=0|TP=0|ip-26-0-161-78]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=97|PP=0|TP=0|ip-26-0-174-196]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=33|PP=0|TP=0|ip-26-0-163-147]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=99|PP=0|TP=0|ip-26-0-174-196]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=98|PP=0|TP=0|ip-26-0-174-196]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=100|PP=0|TP=0|ip-26-0-174-196]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=36|PP=0|TP=0|ip-26-0-163-147]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=34|PP=0|TP=0|ip-26-0-163-147]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=120|PP=0|TP=0|ip-26-0-175-34]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=8|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=104|PP=0|TP=0|ip-26-0-174-240]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=81|PP=0|TP=0|ip-26-0-173-202]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=40|PP=0|TP=0|ip-26-0-163-226]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=17|PP=0|TP=0|ip-26-0-160-242]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=88|PP=0|TP=0|ip-26-0-173-246]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=90|PP=0|TP=0|ip-26-0-173-246]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=113|PP=0|TP=0|ip-26-0-175-19]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=9|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=69|PP=0|TP=0|ip-26-0-170-160]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=72|PP=0|TP=0|ip-26-0-170-31]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=101|PP=0|TP=0|ip-26-0-174-196]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=50|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=121|PP=0|TP=0|ip-26-0-175-34]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=82|PP=0|TP=0|ip-26-0-173-202]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=84|PP=0|TP=0|ip-26-0-173-202]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=57|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=27|PP=0|TP=0|ip-26-0-161-78]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=26|PP=0|TP=0|ip-26-0-161-78]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=18|PP=0|TP=0|ip-26-0-160-242]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=41|PP=0|TP=0|ip-26-0-163-226]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=5|PP=0|TP=0|ip-26-0-160-100]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=19|PP=0|TP=0|ip-26-0-160-242]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=91|PP=0|TP=0|ip-26-0-173-246]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=51|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=105|PP=0|TP=0|ip-26-0-174-240]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=37|PP=0|TP=0|ip-26-0-163-147]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=35|PP=0|TP=0|ip-26-0-163-147]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=83|PP=0|TP=0|ip-26-0-173-202]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=58|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=59|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=112|PP=0|TP=0|ip-26-0-175-19]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=28|PP=0|TP=0|ip-26-0-161-78]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=29|PP=0|TP=0|ip-26-0-161-78]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=20|PP=0|TP=0|ip-26-0-160-242]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=42|PP=0|TP=0|ip-26-0-163-226]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=73|PP=0|TP=0|ip-26-0-170-31]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=92|PP=0|TP=0|ip-26-0-173-246]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=52|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=10|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=60|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=114|PP=0|TP=0|ip-26-0-175-19]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=115|PP=0|TP=0|ip-26-0-175-19]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=30|PP=0|TP=0|ip-26-0-161-78]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=44|PP=0|TP=0|ip-26-0-163-226]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=93|PP=0|TP=0|ip-26-0-173-246]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=38|PP=0|TP=0|ip-26-0-163-147]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=122|PP=0|TP=0|ip-26-0-175-34]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=123|PP=0|TP=0|ip-26-0-175-34]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=85|PP=0|TP=0|ip-26-0-173-202]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=63|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=116|PP=0|TP=0|ip-26-0-175-19]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=31|PP=0|TP=0|ip-26-0-161-78]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=21|PP=0|TP=0|ip-26-0-160-242]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=43|PP=0|TP=0|ip-26-0-163-226]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=74|PP=0|TP=0|ip-26-0-170-31]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=95|PP=0|TP=0|ip-26-0-173-246]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=53|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=12|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=71|PP=0|TP=0|ip-26-0-170-160]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=107|PP=0|TP=0|ip-26-0-174-240]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=108|PP=0|TP=0|ip-26-0-174-240]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=106|PP=0|TP=0|ip-26-0-174-240]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=11|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=70|PP=0|TP=0|ip-26-0-170-160]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=39|PP=0|TP=0|ip-26-0-163-147]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=87|PP=0|TP=0|ip-26-0-173-202]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=7|PP=0|TP=0|ip-26-0-160-100]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=23|PP=0|TP=0|ip-26-0-160-242]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=22|PP=0|TP=0|ip-26-0-160-242]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=75|PP=0|TP=0|ip-26-0-170-31]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=76|PP=0|TP=0|ip-26-0-170-31]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=103|PP=0|TP=0|ip-26-0-174-196]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=102|PP=0|TP=0|ip-26-0-174-196]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=89|PP=0|TP=0|ip-26-0-173-246]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=94|PP=0|TP=0|ip-26-0-173-246]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=124|PP=0|TP=0|ip-26-0-175-34]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=13|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=62|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=117|PP=0|TP=0|ip-26-0-175-19]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=6|PP=0|TP=0|ip-26-0-160-100]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=54|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=55|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=45|PP=0|TP=0|ip-26-0-163-226]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=86|PP=0|TP=0|ip-26-0-173-202]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=125|PP=0|TP=0|ip-26-0-175-34]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=119|PP=0|TP=0|ip-26-0-175-19]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=77|PP=0|TP=0|ip-26-0-170-31]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=61|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=118|PP=0|TP=0|ip-26-0-175-19]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=15|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=47|PP=0|TP=0|ip-26-0-163-226]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=46|PP=0|TP=0|ip-26-0-163-226]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=127|PP=0|TP=0|ip-26-0-175-34]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=78|PP=0|TP=0|ip-26-0-170-31]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=109|PP=0|TP=0|ip-26-0-174-240]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=126|PP=0|TP=0|ip-26-0-175-34]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=14|PP=0|TP=0|ip-26-0-160-225]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=79|PP=0|TP=0|ip-26-0-170-31]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=111|PP=0|TP=0|ip-26-0-174-240]: Throughput logging complete +12/02/2024 10:27:08 [INFO|DP=110|PP=0|TP=0|ip-26-0-174-240]: Throughput logging complete +[2024-12-02 10:27:08,094] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166547 closing signal SIGTERM +srun: Job step aborted: Waiting up to 32 seconds for job step to finish. +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166548 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166549 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166550 closing signal SIGTERM +[2024-12-02 10:27:08,093] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,091] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,091] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2250831 closing signal SIGTERM +[2024-12-02 10:27:08,093] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,093] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3396166 closing signal SIGTERM +[2024-12-02 10:27:08,092] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2250832 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34104 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,092] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2250833 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34105 closing signal SIGTERM +[2024-12-02 10:27:08,093] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3396167 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1243541 closing signal SIGTERM +[2024-12-02 10:27:08,093] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1243542 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3396168 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 291213 closing signal SIGTERM +[2024-12-02 10:27:08,093] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4099551 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1243543 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1243544 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 291214 closing signal SIGTERM +[2024-12-02 10:27:08,093] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4099552 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1525730 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 291215 closing signal SIGTERM +[2024-12-02 10:27:08,093] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4099553 closing signal SIGTERM +[2024-12-02 10:27:08,092] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2250834 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34106 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1525731 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34107 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 291216 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166551 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1525732 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,094] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2654390 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 357190 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2654391 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2956736 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2956737 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 357191 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2654392 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 357192 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 357193 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2956738 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1243545 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2956739 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2654393 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34108 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3396169 closing signal SIGTERM +slurmstepd: error: *** JOB 12178054 ON ip-26-0-160-100 CANCELLED AT 2024-12-02T10:27:08 *** +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3396170 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3209135 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2654394 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1525733 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3209136 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3209137 closing signal SIGTERM +[2024-12-02 10:27:08,093] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2250835 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3396171 closing signal SIGTERM +slurmstepd: error: *** STEP 12178054.0 ON ip-26-0-160-100 CANCELLED AT 2024-12-02T10:27:08 *** +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4099554 closing signal SIGTERM +[2024-12-02 10:27:08,096] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3209138 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4099555 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34109 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 291217 closing signal SIGTERM +[2024-12-02 10:27:08,096] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1525734 closing signal SIGTERM +[2024-12-02 10:27:08,096] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1525735 closing signal SIGTERM +[2024-12-02 10:27:08,096] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1525736 closing signal SIGTERM +[2024-12-02 10:27:08,093] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2250836 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4099556 closing signal SIGTERM +[2024-12-02 10:27:08,095] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4099557 closing signal SIGTERM +[2024-12-02 10:27:08,096] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252225 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2250837 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252226 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,096] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3396172 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1525737 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252227 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2711693 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252228 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2711694 closing signal SIGTERM +[2024-12-02 10:27:08,096] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3396173 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252229 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252230 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2711695 closing signal SIGTERM +[2024-12-02 10:27:08,096] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 357194 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166552 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1243546 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 291218 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 291219 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 291220 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2956740 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2956741 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2956742 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2956743 closing signal SIGTERM +[2024-12-02 10:27:08,096] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34110 closing signal SIGTERM +[2024-12-02 10:27:08,096] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 34111 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4099558 closing signal SIGTERM +[2024-12-02 10:27:08,094] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2250838 closing signal SIGTERM +[2024-12-02 10:27:08,096] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,097] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3209139 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2654395 closing signal SIGTERM +[2024-12-02 10:27:08,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252231 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2711696 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1243547 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1243548 closing signal SIGTERM +[2024-12-02 10:27:08,096] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 357195 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166553 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82894 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40020 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2654396 closing signal SIGTERM +[2024-12-02 10:27:08,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3252232 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 357196 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82895 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40021 closing signal SIGTERM +[2024-12-02 10:27:08,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2654397 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82896 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40022 closing signal SIGTERM +[2024-12-02 10:27:08,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2711697 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 166554 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82897 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40023 closing signal SIGTERM +[2024-12-02 10:27:08,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2711698 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82898 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40024 closing signal SIGTERM +[2024-12-02 10:27:08,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2711699 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82899 closing signal SIGTERM +[2024-12-02 10:27:08,100] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40025 closing signal SIGTERM +[2024-12-02 10:27:08,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2711700 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82900 closing signal SIGTERM +[2024-12-02 10:27:08,099] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 82901 closing signal SIGTERM +[2024-12-02 10:27:08,101] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40026 closing signal SIGTERM +[2024-12-02 10:27:08,097] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 357197 closing signal SIGTERM +[2024-12-02 10:27:08,102] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 40027 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3209140 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3209141 closing signal SIGTERM +[2024-12-02 10:27:08,098] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3209142 closing signal SIGTERM diff --git a/logs/bench_dp64_tp8_pp1_acc1_mbs1_seq2048-11600246.out b/logs/bench_dp64_tp8_pp1_acc1_mbs1_seq2048-11600246.out new file mode 100644 index 0000000000000000000000000000000000000000..bccd817b8bdf10e2ad847899f474ed3e5f5bbbbb --- /dev/null +++ b/logs/bench_dp64_tp8_pp1_acc1_mbs1_seq2048-11600246.out @@ -0,0 +1,1746 @@ ++ source /etc/profile.d/modules.sh +++ . /usr/share/modules/init/bash ++++ unset _mlshdbg ++++ '[' 0 = 1 ']' ++++ unset _mlre _mlIFS ++++ '[' -n x ']' ++++ _mlIFS=' +' ++++ IFS=' ' ++++ '[' -n '' ']' +++++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash autoinit ++++ _mlcode='module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' ++++ _mlret=0 ++++ '[' -n x ']' ++++ IFS=' +' ++++ unset _mlIFS ++++ unset _mlre _mlv _mlrv ++++ '[' 0 -eq 0 ']' ++++ eval 'module() { + unset _mlshdbg; + if [ "${MODULES_SILENT_SHELL_DEBUG:-0}" = '\''1'\'' ]; then + case "$-" in + *v*x*) set +vx; _mlshdbg='\''vx'\'' ;; + *v*) set +v; _mlshdbg='\''v'\'' ;; + *x*) set +x; _mlshdbg='\''x'\'' ;; + *) _mlshdbg='\'''\'' ;; + esac; + fi; + unset _mlre _mlIFS; + if [ -n "${IFS+x}" ]; then + _mlIFS=$IFS; + fi; + IFS='\'' '\''; + for _mlv in ${MODULES_RUN_QUARANTINE:-}; do + if [ "${_mlv}" = "${_mlv##*[!A-Za-z0-9_]}" -a "${_mlv}" = "${_mlv#[0-9]}" ]; then + if [ -n "`eval '\''echo ${'\''$_mlv'\''+x}'\''`" ]; then + _mlre="${_mlre:-}${_mlv}_modquar='\''`eval '\''echo ${'\''$_mlv'\''}'\''`'\'' "; + fi; + _mlrv="MODULES_RUNENV_${_mlv}"; + _mlre="${_mlre:-}${_mlv}='\''`eval '\''echo ${'\''$_mlrv'\'':-}'\''`'\'' "; + fi; + done; + if [ -n "${_mlre:-}" ]; then + eval `eval ${_mlre}/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash '\''"$@"'\''`; + else + eval `/usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash "$@"`; + fi; + _mlstatus=$?; + if [ -n "${_mlIFS+x}" ]; then + IFS=$_mlIFS; + else + unset IFS; + fi; + unset _mlre _mlv _mlrv _mlIFS; + if [ -n "${_mlshdbg:-}" ]; then + set -$_mlshdbg; + fi; + unset _mlshdbg; + return $_mlstatus; +}; +MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl; export MODULES_CMD; +ENV=/usr/share/modules/init/profile.sh; export ENV; +MODULEPATH_modshare=/etc/environment-modules/modules:1:/usr/share/modules/\$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1; export MODULEPATH_modshare; +BASH_ENV=/usr/share/modules/init/bash; export BASH_ENV; +MODULESHOME=/usr/share/modules; export MODULESHOME; +LOADEDMODULES=; export LOADEDMODULES; +MODULEPATH=/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/\$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles; export MODULEPATH; +test 0;' +++++ MODULES_CMD=/usr/lib/x86_64-linux-gnu/modulecmd.tcl +++++ export MODULES_CMD +++++ ENV=/usr/share/modules/init/profile.sh +++++ export ENV +++++ MODULEPATH_modshare='/etc/environment-modules/modules:1:/usr/share/modules/$MODULE_VERSION/modulefiles:1:/usr/share/modules/modulefiles:1:/usr/share/modules/versions:1' +++++ export MODULEPATH_modshare +++++ BASH_ENV=/usr/share/modules/init/bash +++++ export BASH_ENV +++++ MODULESHOME=/usr/share/modules +++++ export MODULESHOME +++++ LOADEDMODULES= +++++ export LOADEDMODULES +++++ MODULEPATH='/etc/environment-modules/modules:/usr/share/modules/versions:/usr/share/modules/$MODULE_VERSION/modulefiles:/usr/share/modules/modulefiles' +++++ export MODULEPATH +++++ test 0 ++++ '[' 0 = 1 ']' ++++ '[' -t 2 ']' ++++ export -f module ++++ export -f switchml ++++ '[' 5 -ge 3 ']' ++++ [[ ehxB =~ i ]] ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin: =~ :/usr/bin: ]] ++++ '[' '!' -n '' ']' ++++ MANPATH=: ++++ export MANPATH +++++ manpath ++++ [[ ! :/fsx/nouamane/miniconda/envs/2-1-cu121/man:/fsx/nouamane/miniconda/envs/2-1-cu121/share/man:/fsx/nouamane/miniconda/man:/fsx/nouamane/miniconda/share/man:/usr/local/man:/usr/local/share/man:/usr/share/man:/opt/slurm/share/man:/admin/home/nouamane/.fzf/man:: =~ :/usr/share/man: ]] ++++ unset _mlcode _mlret ++++ '[' -n '' ']' ++ module load cuda/12.1 ++ unset _mlshdbg ++ '[' 0 = 1 ']' ++ unset _mlre _mlIFS ++ '[' -n x ']' ++ _mlIFS=' +' ++ IFS=' ' ++ '[' -n '' ']' +++ /usr/bin/tclsh8.6 /usr/lib/x86_64-linux-gnu/modulecmd.tcl bash load cuda/12.1 ++ eval 'CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include;' export 'CPATH; +LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib:;' export 'LD_LIBRARY_PATH; +FI_EFA_FORK_SAFE=1;' export 'FI_EFA_FORK_SAFE; +MANPATH=/usr/local/cuda-12.1/share/man::;' export 'MANPATH; +LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64;' export 'LIBRARY_PATH; +_LMFILES_=/usr/share/modules/modulefiles/cuda/12.1;' export '_LMFILES_; +LOADEDMODULES=cuda/12.1;' export 'LOADEDMODULES; +MPI_PATH=/opt/amazon/openmpi;' export 'MPI_PATH; +NCCL_HOME_modshare=/opt/nccl/build:1;' export 'NCCL_HOME_modshare; +NCCL_PROTO=simple;' export 'NCCL_PROTO; +MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1;' export 'MANPATH_modshare; +LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1;' export 'LIBRARY_PATH_modshare; +NCCL_SOCKET_IFNAME=enp;' export 'NCCL_SOCKET_IFNAME; +AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl;' export 'AWS_OFI_NCCL_HOME; +NCCL_HOME=/opt/nccl/build;' export 'NCCL_HOME; +FI_PROVIDER=efa;' export 'FI_PROVIDER; +AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1;' export 'AWS_OFI_NCCL_HOME_modshare; +CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1;' export 'CPATH_modshare; +LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1;' export 'LD_LIBRARY_PATH_modshare; +FI_EFA_ENABLE_SHM_TRANSFER=1;' export 'FI_EFA_ENABLE_SHM_TRANSFER; +_LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1;' export '_LMFILES__modshare; +LOADEDMODULES_modshare=cuda/12.1:1;' export 'LOADEDMODULES_modshare; +MPI_PATH_modshare=/opt/amazon/openmpi:1;' export 'MPI_PATH_modshare; +PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin;' export 'PATH; +CUDA_HOME=/usr/local/cuda-12.1;' export 'CUDA_HOME; +PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/fsx/nouamane/miniconda/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1;' export 'PATH_modshare; +test' '0;' +++ CPATH=/opt/nccl/build/include:/usr/local/cuda-12.1/include +++ export CPATH +++ LD_LIBRARY_PATH=/opt/nccl/build/lib:/opt/aws-ofi-nccl/lib:/opt/amazon/efa/lib:/opt/amazon/openmpi/lib:/usr/local/cuda-12.1/efa/lib:/usr/local/cuda-12.1/lib:/usr/local/cuda-12.1/lib64:/usr/local/cuda-12.1:/usr/local/cuda-12.1/extras/CUPTI/lib64:/usr/local/cuda-12.1/targets/x86_64-linux/lib: +++ export LD_LIBRARY_PATH +++ FI_EFA_FORK_SAFE=1 +++ export FI_EFA_FORK_SAFE +++ MANPATH=/usr/local/cuda-12.1/share/man:: +++ export MANPATH +++ LIBRARY_PATH=/opt/aws-ofi-nccl/lib:/opt/nccl/build/lib:/usr/local/cuda-12.1/lib64 +++ export LIBRARY_PATH +++ _LMFILES_=/usr/share/modules/modulefiles/cuda/12.1 +++ export _LMFILES_ +++ LOADEDMODULES=cuda/12.1 +++ export LOADEDMODULES +++ MPI_PATH=/opt/amazon/openmpi +++ export MPI_PATH +++ NCCL_HOME_modshare=/opt/nccl/build:1 +++ export NCCL_HOME_modshare +++ NCCL_PROTO=simple +++ export NCCL_PROTO +++ MANPATH_modshare=:1:/usr/local/cuda-12.1/share/man:1 +++ export MANPATH_modshare +++ LIBRARY_PATH_modshare=/opt/aws-ofi-nccl/lib:1:/opt/nccl/build/lib:1:/usr/local/cuda-12.1/lib64:1 +++ export LIBRARY_PATH_modshare +++ NCCL_SOCKET_IFNAME=enp +++ export NCCL_SOCKET_IFNAME +++ AWS_OFI_NCCL_HOME=/opt/aws-ofi-nccl +++ export AWS_OFI_NCCL_HOME +++ NCCL_HOME=/opt/nccl/build +++ export NCCL_HOME +++ FI_PROVIDER=efa +++ export FI_PROVIDER +++ AWS_OFI_NCCL_HOME_modshare=/opt/aws-ofi-nccl:1 +++ export AWS_OFI_NCCL_HOME_modshare +++ CPATH_modshare=/usr/local/cuda-12.1/include:1:/opt/nccl/build/include:1 +++ export CPATH_modshare +++ LD_LIBRARY_PATH_modshare=/opt/amazon/efa/lib:3:/opt/amazon/openmpi/lib:2:/usr/local/cuda-12.1/lib:1:/usr/local/cuda-12.1/extras/CUPTI/lib64:2:/usr/local/cuda-12.1/targets/x86_64-linux/lib:2::1:/opt/nccl/build/lib:1:/opt/aws-ofi-nccl/lib:2:/usr/local/cuda-12.1/lib64:2:/usr/local/cuda-12.1:2:/usr/local/cuda-12.1/efa/lib:1 +++ export LD_LIBRARY_PATH_modshare +++ FI_EFA_ENABLE_SHM_TRANSFER=1 +++ export FI_EFA_ENABLE_SHM_TRANSFER +++ _LMFILES__modshare=/usr/share/modules/modulefiles/cuda/12.1:1 +++ export _LMFILES__modshare +++ LOADEDMODULES_modshare=cuda/12.1:1 +++ export LOADEDMODULES_modshare +++ MPI_PATH_modshare=/opt/amazon/openmpi:1 +++ export MPI_PATH_modshare +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export PATH +++ CUDA_HOME=/usr/local/cuda-12.1 +++ export CUDA_HOME +++ PATH_modshare=/usr/local/cuda-12.1/efa/test-cuda-12.1:1:/usr/bin:1:/usr/local/bin:1:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:1:/bin:1:/snap/bin:1:/fsx/nouamane/miniconda/condabin:1:/sbin:1:/opt/slurm/bin:1:/fsx/nouamane/miniconda/bin:1:/usr/sbin:1:/usr/games:1:/usr/local/sbin:1:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:1:/usr/local/cuda-12.1/bin:1:/admin/home/nouamane/.fzf/bin:1:/usr/local/games:1 +++ export PATH_modshare +++ test 0 ++ _mlstatus=0 ++ '[' -n x ']' ++ IFS=' +' ++ unset _mlre _mlv _mlrv _mlIFS ++ '[' -n '' ']' ++ unset _mlshdbg ++ return 0 ++ source /fsx/nouamane/miniconda/bin/activate +++ _CONDA_ROOT=/fsx/nouamane/miniconda +++ . /fsx/nouamane/miniconda/etc/profile.d/conda.sh ++++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda ++++ export _CE_M= ++++ _CE_M= ++++ export _CE_CONDA= ++++ _CE_CONDA= ++++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python ++++ '[' -z x ']' +++ conda activate +++ local cmd=activate +++ case "$cmd" in +++ __conda_activate activate +++ '[' -n '' ']' +++ local ask_conda ++++ PS1= ++++ __conda_exe shell.posix activate ++++ /fsx/nouamane/miniconda/bin/conda shell.posix activate +++ ask_conda='PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' +++ eval 'PS1='\''(base) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_SHLVL='\''1'\'' +export CONDA_PROMPT_MODIFIER='\''(base) '\''' ++++ PS1='(base) ' ++++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++++ export CONDA_SHLVL=1 ++++ CONDA_SHLVL=1 ++++ export 'CONDA_PROMPT_MODIFIER=(base) ' ++++ CONDA_PROMPT_MODIFIER='(base) ' +++ __conda_hashr +++ '[' -n '' ']' +++ '[' -n '' ']' +++ hash -r ++ conda activate 2-1-cu121 ++ local cmd=activate ++ case "$cmd" in ++ __conda_activate activate 2-1-cu121 ++ '[' -n '' ']' ++ local ask_conda +++ PS1='(base) ' +++ __conda_exe shell.posix activate 2-1-cu121 +++ /fsx/nouamane/miniconda/bin/conda shell.posix activate 2-1-cu121 ++ ask_conda='PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' ++ eval 'PS1='\''(2-1-cu121) '\'' +export PATH='\''/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin'\'' +export CONDA_PREFIX='\''/fsx/nouamane/miniconda/envs/2-1-cu121'\'' +export CONDA_SHLVL='\''2'\'' +export CONDA_DEFAULT_ENV='\''2-1-cu121'\'' +export CONDA_PROMPT_MODIFIER='\''(2-1-cu121) '\'' +export CONDA_PREFIX_1='\''/fsx/nouamane/miniconda'\'' +export CONDA_EXE='\''/fsx/nouamane/miniconda/bin/conda'\'' +export _CE_M='\'''\'' +export _CE_CONDA='\'''\'' +export CONDA_PYTHON_EXE='\''/fsx/nouamane/miniconda/bin/python'\'' +. "/fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh"' +++ PS1='(2-1-cu121) ' +++ export PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ PATH=/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ export CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ CONDA_PREFIX=/fsx/nouamane/miniconda/envs/2-1-cu121 +++ export CONDA_SHLVL=2 +++ CONDA_SHLVL=2 +++ export CONDA_DEFAULT_ENV=2-1-cu121 +++ CONDA_DEFAULT_ENV=2-1-cu121 +++ export 'CONDA_PROMPT_MODIFIER=(2-1-cu121) ' +++ CONDA_PROMPT_MODIFIER='(2-1-cu121) ' +++ export CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ CONDA_PREFIX_1=/fsx/nouamane/miniconda +++ export CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ CONDA_EXE=/fsx/nouamane/miniconda/bin/conda +++ export _CE_M= +++ _CE_M= +++ export _CE_CONDA= +++ _CE_CONDA= +++ export CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ CONDA_PYTHON_EXE=/fsx/nouamane/miniconda/bin/python +++ . /fsx/nouamane/miniconda/envs/2-1-cu121/etc/conda/activate.d/libxml2_activate.sh ++++ test -n '' ++++ xml_catalog_files_libxml2= ++++ XML_CATALOG_FILES= ++++ conda_catalog_files= ++++ ifs_libxml2=' +' ++++ IFS=' ' ++++ rem=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ for pre in ${rem} ++++ test '' = /fsx/nouamane/miniconda/envs/2-1-cu121 ++++ conda_catalog_files=/fsx/nouamane/miniconda/envs/2-1-cu121 ++++ rem= ++++ IFS=' +' ++++ conda_catalog_files='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ export 'XML_CATALOG_FILES=file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ XML_CATALOG_FILES='file:///fsx/nouamane/miniconda/envs/2-1-cu121/etc/xml/catalog file:///etc/xml/catalog' ++++ unset conda_catalog_files ifs_libxml2 rem ++ __conda_hashr ++ '[' -n '' ']' ++ '[' -n '' ']' ++ hash -r ++ export PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin ++ PATH=/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/usr/local/cuda-12.1/efa/test-cuda-12.1:/usr/local/cuda-12.1/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/envs/2-1-cu121/bin:/fsx/nouamane/miniconda/condabin:/admin/home/nouamane/.cursor-server/bin/001668006cc714afd397f4ef0d52862f5a095530/bin/remote-cli:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/opt/slurm/bin:/admin/home/nouamane/.fzf/bin +++ scontrol show hostnames 'ip-26-0-160-103,ip-26-0-161-[78,123],ip-26-0-162-[46,233],ip-26-0-163-[43,158,220,236],ip-26-0-164-[0,18,45,75,187,236],ip-26-0-165-[24,38,59,164,202,213],ip-26-0-166-[15,214,244],ip-26-0-167-[9,51,175,177],ip-26-0-168-[34,95,120],ip-26-0-169-[132,139,207,247],ip-26-0-170-[31,132,143,160],ip-26-0-171-[21,56,62,88,168,230,249],ip-26-0-172-[57,73,116,142,147],ip-26-0-173-[7,202,246],ip-26-0-174-[100,186,196,240],ip-26-0-175-[19,34,132,165,170,241]' ++ export 'NODELIST=ip-26-0-160-103 +ip-26-0-161-78 +ip-26-0-161-123 +ip-26-0-162-46 +ip-26-0-162-233 +ip-26-0-163-43 +ip-26-0-163-158 +ip-26-0-163-220 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-168-34 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-173-7 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165 +ip-26-0-175-170 +ip-26-0-175-241' ++ NODELIST='ip-26-0-160-103 +ip-26-0-161-78 +ip-26-0-161-123 +ip-26-0-162-46 +ip-26-0-162-233 +ip-26-0-163-43 +ip-26-0-163-158 +ip-26-0-163-220 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-168-34 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-173-7 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165 +ip-26-0-175-170 +ip-26-0-175-241' +++ head -n1 +++ scontrol show hostnames 'ip-26-0-160-103,ip-26-0-161-[78,123],ip-26-0-162-[46,233],ip-26-0-163-[43,158,220,236],ip-26-0-164-[0,18,45,75,187,236],ip-26-0-165-[24,38,59,164,202,213],ip-26-0-166-[15,214,244],ip-26-0-167-[9,51,175,177],ip-26-0-168-[34,95,120],ip-26-0-169-[132,139,207,247],ip-26-0-170-[31,132,143,160],ip-26-0-171-[21,56,62,88,168,230,249],ip-26-0-172-[57,73,116,142,147],ip-26-0-173-[7,202,246],ip-26-0-174-[100,186,196,240],ip-26-0-175-[19,34,132,165,170,241]' ++ export MASTER_NODE=ip-26-0-160-103 ++ MASTER_NODE=ip-26-0-160-103 ++ export MASTER_PORT=12356 ++ MASTER_PORT=12356 ++ export NNODES=64 ++ NNODES=64 ++ export GPUS_PER_NODE=8 ++ GPUS_PER_NODE=8 ++ export WORLD_SIZE=512 ++ WORLD_SIZE=512 ++ export CUDA_DEVICE_MAX_CONNECTIONS=1 ++ CUDA_DEVICE_MAX_CONNECTIONS=1 ++ export NANOTRON_BENCHMARK=1 ++ NANOTRON_BENCHMARK=1 ++ echo 'Master node: ip-26-0-160-103' +Master node: ip-26-0-160-103 ++ echo 'All nodes: ip-26-0-160-103 +ip-26-0-161-78 +ip-26-0-161-123 +ip-26-0-162-46 +ip-26-0-162-233 +ip-26-0-163-43 +ip-26-0-163-158 +ip-26-0-163-220 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-168-34 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-173-7 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165 +ip-26-0-175-170 +ip-26-0-175-241' +All nodes: ip-26-0-160-103 +ip-26-0-161-78 +ip-26-0-161-123 +ip-26-0-162-46 +ip-26-0-162-233 +ip-26-0-163-43 +ip-26-0-163-158 +ip-26-0-163-220 +ip-26-0-163-236 +ip-26-0-164-0 +ip-26-0-164-18 +ip-26-0-164-45 +ip-26-0-164-75 +ip-26-0-164-187 +ip-26-0-164-236 +ip-26-0-165-24 +ip-26-0-165-38 +ip-26-0-165-59 +ip-26-0-165-164 +ip-26-0-165-202 +ip-26-0-165-213 +ip-26-0-166-15 +ip-26-0-166-214 +ip-26-0-166-244 +ip-26-0-167-9 +ip-26-0-167-51 +ip-26-0-167-175 +ip-26-0-167-177 +ip-26-0-168-34 +ip-26-0-168-95 +ip-26-0-168-120 +ip-26-0-169-132 +ip-26-0-169-139 +ip-26-0-169-207 +ip-26-0-169-247 +ip-26-0-170-31 +ip-26-0-170-132 +ip-26-0-170-143 +ip-26-0-170-160 +ip-26-0-171-21 +ip-26-0-171-56 +ip-26-0-171-62 +ip-26-0-171-88 +ip-26-0-171-168 +ip-26-0-171-230 +ip-26-0-171-249 +ip-26-0-172-57 +ip-26-0-172-73 +ip-26-0-172-116 +ip-26-0-172-142 +ip-26-0-172-147 +ip-26-0-173-7 +ip-26-0-173-202 +ip-26-0-173-246 +ip-26-0-174-100 +ip-26-0-174-186 +ip-26-0-174-196 +ip-26-0-174-240 +ip-26-0-175-19 +ip-26-0-175-34 +ip-26-0-175-132 +ip-26-0-175-165 +ip-26-0-175-170 +ip-26-0-175-241 ++ echo 'World size: 512' +World size: 512 ++ srun torchrun --nnodes=64 --nproc_per_node=8 --rdzv_id=11600246 --rdzv_backend=c10d --rdzv_endpoint=ip-26-0-160-103:12356 run_train.py --config-file benchmark/configs/config_dp64_tp8_pp1_acc1_mbs1_seq2048.yaml +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,203] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,203] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,203] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,203] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,203] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,203] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,204] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,205] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,205] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,205] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,205] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,205] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,206] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,206] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,206] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,206] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,206] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,208] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,208] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,208] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,208] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,208] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,207] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,210] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,212] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,212] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,212] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,212] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,212] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,217] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,216] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,217] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,217] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,217] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,217] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,216] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,216] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,216] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,216] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,222] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,223] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,223] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,223] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,223] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,227] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,227] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,227] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,227] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,227] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,229] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,229] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,229] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,229] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,229] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,230] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,230] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,230] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,230] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,230] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,229] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,229] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,229] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,229] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,229] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,233] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,233] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,233] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,233] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,233] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,234] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,234] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,234] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,234] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,234] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,235] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,235] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,235] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,235] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,235] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,236] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,236] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,236] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,236] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,236] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,237] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,237] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,237] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,237] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,237] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,238] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,239] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,239] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,239] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,239] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,238] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,239] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,238] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,238] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,238] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,238] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,239] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,239] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,239] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,239] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,240] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,241] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,241] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,241] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,241] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,241] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,241] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,241] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,241] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,241] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,243] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,244] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,244] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,244] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,244] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,244] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,244] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,244] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,244] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,244] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,248] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,248] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,248] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,248] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,248] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,249] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,249] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,249] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,249] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,249] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,251] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,251] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,251] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,251] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,251] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,253] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,253] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,253] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,253] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,253] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,256] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,256] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,256] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,256] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,256] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,257] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,257] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,257] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,257] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,257] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,260] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,260] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,260] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,260] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,260] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,262] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,262] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,262] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,262] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,262] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,266] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,266] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,266] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,266] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,266] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,269] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,269] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,269] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,269] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,269] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,268] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,268] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,268] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,268] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,268] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,272] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,272] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,272] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,272] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,272] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,273] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,273] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,273] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,273] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,273] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,276] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,276] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,276] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,276] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,276] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,291] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,291] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,291] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,291] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,291] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,292] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,293] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,293] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,293] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,293] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,301] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,301] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,301] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,301] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,301] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,312] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,312] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,312] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,312] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,312] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,320] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,320] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,320] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,320] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,320] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,320] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,320] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,320] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,320] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,320] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,333] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,333] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,333] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,333] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,333] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,338] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,338] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,338] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,338] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,338] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,376] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:47,376] torch.distributed.run: [WARNING] +[2024-11-27 14:24:47,376] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:47,376] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:47,376] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:50,093] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:50,093] torch.distributed.run: [WARNING] +[2024-11-27 14:24:50,093] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:50,093] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:50,093] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:50,233] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:50,233] torch.distributed.run: [WARNING] +[2024-11-27 14:24:50,233] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:50,233] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:50,233] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:50,484] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:50,484] torch.distributed.run: [WARNING] +[2024-11-27 14:24:50,484] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:50,484] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:50,484] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:50,534] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:50,534] torch.distributed.run: [WARNING] +[2024-11-27 14:24:50,534] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:50,534] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:50,534] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:50,789] torch.distributed.run: [WARNING] master_addr is only used for static rdzv_backend and when rdzv_endpoint is not specified. +[2024-11-27 14:24:50,789] torch.distributed.run: [WARNING] +[2024-11-27 14:24:50,789] torch.distributed.run: [WARNING] ***************************************** +[2024-11-27 14:24:50,789] torch.distributed.run: [WARNING] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +[2024-11-27 14:24:50,789] torch.distributed.run: [WARNING] ***************************************** +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Config: +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Config(general=GeneralArgs(project='debug', +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: run='dp64_tp8_pp1_acc1_mbs1_seq2048', +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: seed=42, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: step=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: consumed_train_samples=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: benchmark_csv_path=PosixPath('bench.csv'), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: ignore_sanity_checks=True), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: parallelism=ParallelismArgs(dp=64, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pp=1, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp=8, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pp_engine=, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_mode=, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_linear_async_communication=True, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: recompute_layer=False, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tp_recompute_allgather=True, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: expert_parallel_size=1), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: model=ModelArgs(model_config=LlamaConfig(bos_token_id=0, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: eos_token_id=0, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_act='silu', +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_size=2048, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: initializer_range=0.02, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: intermediate_size=8192, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: is_llama_config=True, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: max_position_embeddings=2048, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_attention_heads=32, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_hidden_layers=24, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_key_value_heads=32, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pad_token_id=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pretraining_tp=1, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rms_norm_eps=1e-05, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_scaling=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_theta=10000.0, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_interleaved=False, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tie_word_embeddings=True, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: use_cache=True, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: vocab_size=49152), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: init_method=RandomInit(std=0.02), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: dtype=torch.bfloat16, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: make_vocab_size_divisible_by=1, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: ddp_bucket_cap_mb=25), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer=TokenizerArgs(tokenizer_name_or_path='robot-test/dummy-tokenizer-wordlevel', +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer_revision=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokenizer_max_length=None), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoints=CheckpointsArgs(checkpoints_path=PosixPath('checkpoints'), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoint_interval=10000, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: save_initial_state=False, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: save_final_state=False, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: resume_checkpoint_path=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: checkpoints_path_is_shared_file_system=False), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: logging=LoggingArgs(log_level='info', +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: log_level_replica='info', +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration_step_info_interval=1), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tokens=TokensArgs(sequence_length=2048, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: train_steps=100, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: micro_batch_size=1, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: batch_accumulation_per_replica=1, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: val_check_interval=100, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: limit_val_batches=0, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: limit_test_batches=0), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: optimizer=OptimizerArgs(optimizer_factory=AdamWOptimizerArgs(adam_eps=1e-08, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: adam_beta1=0.9, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: adam_beta2=0.95, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: torch_adam_is_fused=True, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: name='adamW'), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: zero_stage=0, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: weight_decay=0.01, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: clip_grad=1.0, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: accumulate_grad_in_fp32=True, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: learning_rate_scheduler=LRSchedulerArgs(learning_rate=0.0003, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_warmup_steps=2, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_warmup_style='linear', +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_style='cosine', +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_steps=13, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lr_decay_starting_step=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: min_decay_lr=1e-05)), +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: data_stages=[DatasetStageArgs(name='Stable Training Stage', +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: start_training_step=1, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: data=DataArgs(dataset=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: seed=42, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_loading_workers=1))], +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: profiler=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: lighteval=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: s3_upload=None) +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Model Config: +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: LlamaConfig(bos_token_id=0, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: eos_token_id=0, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_act='silu', +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: hidden_size=2048, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: initializer_range=0.02, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: intermediate_size=8192, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: is_llama_config=True, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: max_position_embeddings=2048, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_attention_heads=32, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_hidden_layers=24, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: num_key_value_heads=32, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pad_token_id=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: pretraining_tp=1, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rms_norm_eps=1e-05, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_scaling=None, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_theta=10000.0, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: rope_interleaved=False, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: tie_word_embeddings=True, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: use_cache=True, +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: vocab_size=49152) +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Building model.. +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Initialize RoPE Theta = 10000.0 +11/27/2024 14:25:49 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Setting PP block ranks... +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=5|ip-26-0-160-103]: Local number of parameters: 214M (408.19MiB) +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=7|ip-26-0-160-103]: Local number of parameters: 214M (408.19MiB) +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Total number of parameters: 1.71G (3265.53MiB) +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-103]: Local number of parameters: 214M (408.19MiB) +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=6|ip-26-0-160-103]: Local number of parameters: 214M (408.19MiB) +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Local number of parameters: 214M (408.19MiB) +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=4|ip-26-0-160-103]: Local number of parameters: 214M (408.19MiB) +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-103]: Local number of parameters: 214M (408.19MiB) +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=5|ip-26-0-160-103]: [After model building] Memory usage: 409.22MiB. Peak allocated: 5380.00MiB Peak reserved: 13614.00MiB +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=7|ip-26-0-160-103]: [After model building] Memory usage: 409.22MiB. Peak allocated: 5380.00MiB Peak reserved: 13614.00MiB +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-103]: [After model building] Memory usage: 409.22MiB. Peak allocated: 5380.00MiB Peak reserved: 13614.00MiB +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=6|ip-26-0-160-103]: [After model building] Memory usage: 409.22MiB. Peak allocated: 5380.00MiB Peak reserved: 13614.00MiB +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [After model building] Memory usage: 409.22MiB. Peak allocated: 5380.00MiB Peak reserved: 13614.00MiB +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-103]: [After model building] Memory usage: 409.22MiB. Peak allocated: 5380.00MiB Peak reserved: 13614.00MiB +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=4|ip-26-0-160-103]: [After model building] Memory usage: 409.22MiB. Peak allocated: 5380.00MiB Peak reserved: 13614.00MiB +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-103]: Local number of parameters: 214M (408.19MiB) +11/27/2024 14:25:51 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-103]: [After model building] Memory usage: 409.22MiB. Peak allocated: 5380.00MiB Peak reserved: 13614.00MiB +11/27/2024 14:25:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: No checkpoint path provided. +11/27/2024 14:25:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Parametrizing model parameters using StandardParametrizator +11/27/2024 14:25:53 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Optimizer Building] Using LearningRateForSP as learning rate +11/27/2024 14:25:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Training Plan] Stage Stable Training Stage has 99 remaining training steps and has consumed 0 samples +11/27/2024 14:25:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Using dummy data generator +11/27/2024 14:25:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Training Plan] There are 1 training stages +11/27/2024 14:25:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Stage Stable Training Stage] start from step 1 +11/27/2024 14:25:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: +11/27/2024 14:25:54 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: [Start training] datetime: 2024-11-27 14:25:54.355122 | mbs: 1 | grad_accum: 1 | global_batch_size: 64 | sequence_length: 2048 | train_steps: 100 | start_iteration_step: 0 | consumed_train_samples: 0 +wandb: Tracking run with wandb version 0.16.0 +wandb: W&B syncing is set to `offline` in this directory. +wandb: Run `wandb online` or set WANDB_MODE=online to enable cloud syncing. +11/27/2024 14:26:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Resuming training from stage Stable Training Stage, it has trained for 0 samples and has 99 remaining train steps +11/27/2024 14:26:02 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 2450.18MiB. Peak allocated 5380.00MiB. Peak reserved: 13614.00MiB +11/27/2024 14:26:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 2536.22MiB. Peak allocated 3431.06MiB. Peak reserved: 3506.00MiB +11/27/2024 14:26:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration: 1 / 100 | consumed_tokens: 131K | elapsed_time_per_iteration_ms: 2.43K | tokens_per_sec: 53.9K | tokens_per_sec_per_gpu: 105 | global_batch_size: 64 | lm_loss: 11.2 | lr: 0.00015 | model_tflops_per_gpu: 1.21 | hardware_tflops_per_gpu: 1.21 | grad_norm: 1.95 | cuda_memory_allocated: 4.37G | cuda_max_memory_reserved: 4.52G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.3G | hd_free_memory_tb: 243G +11/27/2024 14:26:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4170.25MiB. Peak allocated 4170.25MiB. Peak reserved: 4306.00MiB +11/27/2024 14:26:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4170.25MiB. Peak allocated 4905.17MiB. Peak reserved: 5072.00MiB +11/27/2024 14:26:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration: 2 / 100 | consumed_tokens: 262K | elapsed_time_per_iteration_ms: 396 | tokens_per_sec: 331K | tokens_per_sec_per_gpu: 647 | global_batch_size: 64 | lm_loss: 11.2 | lr: 0.0003 | model_tflops_per_gpu: 7.42 | hardware_tflops_per_gpu: 7.42 | grad_norm: 1.95 | cuda_memory_allocated: 4.37G | cuda_max_memory_reserved: 5.32G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.3G | hd_free_memory_tb: 243G +11/27/2024 14:26:04 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4170.25MiB. Peak allocated 4170.32MiB. Peak reserved: 5072.00MiB +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Memory usage: 4170.25MiB. Peak allocated 4904.17MiB. Peak reserved: 5072.00MiB +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: iteration: 3 / 100 | consumed_tokens: 393K | elapsed_time_per_iteration_ms: 190 | tokens_per_sec: 689K | tokens_per_sec_per_gpu: 1.35K | global_batch_size: 64 | lm_loss: 11.2 | lr: 0.000296 | model_tflops_per_gpu: 15.4 | hardware_tflops_per_gpu: 15.4 | grad_norm: 1.94 | cuda_memory_allocated: 4.37G | cuda_max_memory_reserved: 5.32G | hd_total_memory_tb: 312G | hd_used_memory_tb: 69.3G | hd_free_memory_tb: 243G +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: | job_id | name | nodes | seq_len | mbs | batch_accum | gbs | mTFLOPs | hTFLOPs | tok/s/gpu | AllReduce (GB/s) | AllGather (GB/s) | ReduceScatter (GB/s) | AR Intra-node (GB/s) | AG Intra-node (GB/s) | RS Intra-node (GB/s) | Mem Alloc (GB) | Mem Res (GB) | +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: | -------- | ------------------------------ | ----- | ------- | --- | ----------- | --- | ------- | ------- | --------- | ---------------- | ---------------- | -------------------- | -------------------- | -------------------- | -------------------- | -------------- | ------------ | +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: | 11600246 | dp64_tp8_pp1_acc1_mbs1_seq2048 | 64 | 2048 | 1 | 1 | 64 | 15.44 | 15.44 | 1345.12 | 169.26 | 40.43 | 43.13 | 451.14 | 259.09 | 261.70 | 4.07 | 4.95 | +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=0|ip-26-0-160-103]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=7|ip-26-0-160-103]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=6|ip-26-0-160-103]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=1|ip-26-0-160-103]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=16|PP=0|TP=5|ip-26-0-165-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=16|PP=0|TP=4|ip-26-0-165-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=48|PP=0|TP=3|ip-26-0-172-147]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=48|PP=0|TP=4|ip-26-0-172-147]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=32|PP=0|TP=4|ip-26-0-169-139]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=32|PP=0|TP=6|ip-26-0-169-139]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=32|PP=0|TP=0|ip-26-0-169-139]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=32|PP=0|TP=2|ip-26-0-169-139]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=5|ip-26-0-160-103]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=2|ip-26-0-160-103]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=4|ip-26-0-160-103]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=40|PP=0|TP=5|ip-26-0-171-21]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=40|PP=0|TP=7|ip-26-0-171-21]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=40|PP=0|TP=1|ip-26-0-171-21]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=8|PP=0|TP=3|ip-26-0-163-43]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=8|PP=0|TP=4|ip-26-0-163-43]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=42|PP=0|TP=3|ip-26-0-171-249]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=42|PP=0|TP=4|ip-26-0-171-249]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=16|PP=0|TP=7|ip-26-0-165-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=16|PP=0|TP=0|ip-26-0-165-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=16|PP=0|TP=2|ip-26-0-165-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=16|PP=0|TP=1|ip-26-0-165-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=16|PP=0|TP=6|ip-26-0-165-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=52|PP=0|TP=1|ip-26-0-173-246]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=44|PP=0|TP=1|ip-26-0-171-62]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=24|PP=0|TP=4|ip-26-0-167-175]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=24|PP=0|TP=2|ip-26-0-167-175]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=56|PP=0|TP=0|ip-26-0-174-196]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=56|PP=0|TP=5|ip-26-0-174-196]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=56|PP=0|TP=6|ip-26-0-174-196]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=56|PP=0|TP=4|ip-26-0-174-196]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=56|PP=0|TP=2|ip-26-0-174-196]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=56|PP=0|TP=7|ip-26-0-174-196]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=56|PP=0|TP=3|ip-26-0-174-196]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=56|PP=0|TP=1|ip-26-0-174-196]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=60|PP=0|TP=7|ip-26-0-175-170]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=60|PP=0|TP=1|ip-26-0-175-170]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=60|PP=0|TP=6|ip-26-0-175-170]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=48|PP=0|TP=2|ip-26-0-172-147]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=48|PP=0|TP=7|ip-26-0-172-147]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=48|PP=0|TP=0|ip-26-0-172-147]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=32|PP=0|TP=3|ip-26-0-169-139]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=32|PP=0|TP=7|ip-26-0-169-139]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=32|PP=0|TP=5|ip-26-0-169-139]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=28|PP=0|TP=6|ip-26-0-168-120]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=28|PP=0|TP=5|ip-26-0-168-120]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=28|PP=0|TP=1|ip-26-0-168-120]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=28|PP=0|TP=2|ip-26-0-168-120]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=0|PP=0|TP=3|ip-26-0-160-103]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=6|PP=0|TP=4|ip-26-0-163-220]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=6|PP=0|TP=5|ip-26-0-163-220]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=40|PP=0|TP=4|ip-26-0-171-21]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=40|PP=0|TP=6|ip-26-0-171-21]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=40|PP=0|TP=2|ip-26-0-171-21]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=8|PP=0|TP=6|ip-26-0-163-43]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=8|PP=0|TP=1|ip-26-0-163-43]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=8|PP=0|TP=2|ip-26-0-163-43]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=25|PP=0|TP=6|ip-26-0-167-177]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=42|PP=0|TP=0|ip-26-0-171-249]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=42|PP=0|TP=1|ip-26-0-171-249]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=16|PP=0|TP=3|ip-26-0-165-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=31|PP=0|TP=6|ip-26-0-169-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=31|PP=0|TP=1|ip-26-0-169-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=13|PP=0|TP=5|ip-26-0-164-45]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=18|PP=0|TP=1|ip-26-0-165-24]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=18|PP=0|TP=4|ip-26-0-165-24]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=47|PP=0|TP=5|ip-26-0-172-142]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=47|PP=0|TP=2|ip-26-0-172-142]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=47|PP=0|TP=0|ip-26-0-172-142]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=47|PP=0|TP=3|ip-26-0-172-142]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=52|PP=0|TP=4|ip-26-0-173-246]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=52|PP=0|TP=3|ip-26-0-173-246]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=52|PP=0|TP=5|ip-26-0-173-246]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=52|PP=0|TP=7|ip-26-0-173-246]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=52|PP=0|TP=6|ip-26-0-173-246]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=57|PP=0|TP=2|ip-26-0-174-240]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=20|PP=0|TP=6|ip-26-0-165-59]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=20|PP=0|TP=2|ip-26-0-165-59]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=20|PP=0|TP=0|ip-26-0-165-59]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=20|PP=0|TP=7|ip-26-0-165-59]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=20|PP=0|TP=3|ip-26-0-165-59]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=39|PP=0|TP=4|ip-26-0-171-168]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=39|PP=0|TP=1|ip-26-0-171-168]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=58|PP=0|TP=0|ip-26-0-175-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=58|PP=0|TP=4|ip-26-0-175-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=58|PP=0|TP=2|ip-26-0-175-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=58|PP=0|TP=5|ip-26-0-175-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=58|PP=0|TP=6|ip-26-0-175-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=24|PP=0|TP=3|ip-26-0-167-175]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=61|PP=0|TP=7|ip-26-0-175-19]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=15|PP=0|TP=2|ip-26-0-165-164]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=15|PP=0|TP=6|ip-26-0-165-164]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=7|PP=0|TP=0|ip-26-0-163-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=7|PP=0|TP=1|ip-26-0-163-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=54|PP=0|TP=1|ip-26-0-174-100]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=14|PP=0|TP=2|ip-26-0-164-75]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=41|PP=0|TP=1|ip-26-0-171-230]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=17|PP=0|TP=0|ip-26-0-165-213]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=17|PP=0|TP=4|ip-26-0-165-213]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=17|PP=0|TP=1|ip-26-0-165-213]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=17|PP=0|TP=2|ip-26-0-165-213]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=22|PP=0|TP=5|ip-26-0-166-214]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=22|PP=0|TP=6|ip-26-0-166-214]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=22|PP=0|TP=1|ip-26-0-166-214]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=5|PP=0|TP=6|ip-26-0-163-158]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=43|PP=0|TP=1|ip-26-0-171-56]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=43|PP=0|TP=7|ip-26-0-171-56]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=43|PP=0|TP=6|ip-26-0-171-56]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=49|PP=0|TP=1|ip-26-0-172-57]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=49|PP=0|TP=5|ip-26-0-172-57]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=49|PP=0|TP=7|ip-26-0-172-57]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=49|PP=0|TP=0|ip-26-0-172-57]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=49|PP=0|TP=3|ip-26-0-172-57]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=29|PP=0|TP=2|ip-26-0-168-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=9|PP=0|TP=3|ip-26-0-164-0]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=9|PP=0|TP=2|ip-26-0-164-0]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=2|PP=0|TP=7|ip-26-0-161-78]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=26|PP=0|TP=6|ip-26-0-167-51]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=26|PP=0|TP=0|ip-26-0-167-51]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=26|PP=0|TP=2|ip-26-0-167-51]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=23|PP=0|TP=6|ip-26-0-166-244]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=23|PP=0|TP=3|ip-26-0-166-244]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=21|PP=0|TP=1|ip-26-0-166-15]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=21|PP=0|TP=7|ip-26-0-166-15]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=59|PP=0|TP=5|ip-26-0-175-165]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=59|PP=0|TP=7|ip-26-0-175-165]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=4|PP=0|TP=0|ip-26-0-162-46]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=4|PP=0|TP=2|ip-26-0-162-46]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=4|PP=0|TP=1|ip-26-0-162-46]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=4|PP=0|TP=4|ip-26-0-162-46]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=4|PP=0|TP=3|ip-26-0-162-46]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=4|PP=0|TP=6|ip-26-0-162-46]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=4|PP=0|TP=7|ip-26-0-162-46]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=63|PP=0|TP=3|ip-26-0-175-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=63|PP=0|TP=7|ip-26-0-175-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=30|PP=0|TP=4|ip-26-0-168-95]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=30|PP=0|TP=2|ip-26-0-168-95]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=30|PP=0|TP=7|ip-26-0-168-95]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=30|PP=0|TP=6|ip-26-0-168-95]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=11|PP=0|TP=1|ip-26-0-164-187]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=11|PP=0|TP=6|ip-26-0-164-187]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=11|PP=0|TP=5|ip-26-0-164-187]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=11|PP=0|TP=4|ip-26-0-164-187]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=11|PP=0|TP=3|ip-26-0-164-187]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=38|PP=0|TP=6|ip-26-0-170-31]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=38|PP=0|TP=4|ip-26-0-170-31]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=38|PP=0|TP=2|ip-26-0-170-31]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=53|PP=0|TP=2|ip-26-0-173-7]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=53|PP=0|TP=4|ip-26-0-173-7]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=33|PP=0|TP=0|ip-26-0-169-207]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=33|PP=0|TP=2|ip-26-0-169-207]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=33|PP=0|TP=3|ip-26-0-169-207]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=33|PP=0|TP=5|ip-26-0-169-207]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=33|PP=0|TP=7|ip-26-0-169-207]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=33|PP=0|TP=4|ip-26-0-169-207]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=50|PP=0|TP=3|ip-26-0-172-73]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=50|PP=0|TP=4|ip-26-0-172-73]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=62|PP=0|TP=3|ip-26-0-175-241]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=46|PP=0|TP=1|ip-26-0-172-116]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=46|PP=0|TP=4|ip-26-0-172-116]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=60|PP=0|TP=5|ip-26-0-175-170]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=60|PP=0|TP=3|ip-26-0-175-170]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=60|PP=0|TP=4|ip-26-0-175-170]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=48|PP=0|TP=6|ip-26-0-172-147]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=48|PP=0|TP=5|ip-26-0-172-147]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=37|PP=0|TP=5|ip-26-0-170-160]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=37|PP=0|TP=3|ip-26-0-170-160]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=37|PP=0|TP=7|ip-26-0-170-160]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=1|PP=0|TP=5|ip-26-0-161-123]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=32|PP=0|TP=1|ip-26-0-169-139]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=12|PP=0|TP=5|ip-26-0-164-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=12|PP=0|TP=4|ip-26-0-164-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=12|PP=0|TP=2|ip-26-0-164-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=12|PP=0|TP=3|ip-26-0-164-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=45|PP=0|TP=6|ip-26-0-171-88]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=45|PP=0|TP=7|ip-26-0-171-88]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=10|PP=0|TP=3|ip-26-0-164-18]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=28|PP=0|TP=0|ip-26-0-168-120]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=19|PP=0|TP=5|ip-26-0-165-38]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=34|PP=0|TP=4|ip-26-0-169-247]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=27|PP=0|TP=3|ip-26-0-167-9]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=27|PP=0|TP=6|ip-26-0-167-9]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=35|PP=0|TP=7|ip-26-0-170-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=35|PP=0|TP=0|ip-26-0-170-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=35|PP=0|TP=6|ip-26-0-170-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=35|PP=0|TP=4|ip-26-0-170-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=35|PP=0|TP=5|ip-26-0-170-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=51|PP=0|TP=1|ip-26-0-173-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=3|PP=0|TP=4|ip-26-0-162-233]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=55|PP=0|TP=6|ip-26-0-174-186]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=55|PP=0|TP=3|ip-26-0-174-186]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=36|PP=0|TP=4|ip-26-0-170-143]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=36|PP=0|TP=1|ip-26-0-170-143]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=36|PP=0|TP=5|ip-26-0-170-143]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=36|PP=0|TP=6|ip-26-0-170-143]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=36|PP=0|TP=7|ip-26-0-170-143]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=40|PP=0|TP=3|ip-26-0-171-21]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=8|PP=0|TP=0|ip-26-0-163-43]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=25|PP=0|TP=4|ip-26-0-167-177]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=25|PP=0|TP=0|ip-26-0-167-177]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=25|PP=0|TP=3|ip-26-0-167-177]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=25|PP=0|TP=5|ip-26-0-167-177]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=25|PP=0|TP=7|ip-26-0-167-177]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=42|PP=0|TP=5|ip-26-0-171-249]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=31|PP=0|TP=2|ip-26-0-169-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=31|PP=0|TP=4|ip-26-0-169-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=31|PP=0|TP=5|ip-26-0-169-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=13|PP=0|TP=3|ip-26-0-164-45]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=13|PP=0|TP=1|ip-26-0-164-45]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=13|PP=0|TP=2|ip-26-0-164-45]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=18|PP=0|TP=0|ip-26-0-165-24]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=47|PP=0|TP=4|ip-26-0-172-142]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=47|PP=0|TP=6|ip-26-0-172-142]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=47|PP=0|TP=7|ip-26-0-172-142]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=47|PP=0|TP=1|ip-26-0-172-142]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=52|PP=0|TP=0|ip-26-0-173-246]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=57|PP=0|TP=5|ip-26-0-174-240]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=57|PP=0|TP=3|ip-26-0-174-240]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=57|PP=0|TP=7|ip-26-0-174-240]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=20|PP=0|TP=4|ip-26-0-165-59]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=20|PP=0|TP=5|ip-26-0-165-59]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=39|PP=0|TP=5|ip-26-0-171-168]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=39|PP=0|TP=6|ip-26-0-171-168]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=39|PP=0|TP=3|ip-26-0-171-168]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=58|PP=0|TP=3|ip-26-0-175-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=24|PP=0|TP=6|ip-26-0-167-175]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=24|PP=0|TP=5|ip-26-0-167-175]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=24|PP=0|TP=7|ip-26-0-167-175]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=61|PP=0|TP=1|ip-26-0-175-19]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=15|PP=0|TP=0|ip-26-0-165-164]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=15|PP=0|TP=1|ip-26-0-165-164]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=15|PP=0|TP=3|ip-26-0-165-164]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=7|PP=0|TP=6|ip-26-0-163-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=7|PP=0|TP=4|ip-26-0-163-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=54|PP=0|TP=0|ip-26-0-174-100]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=54|PP=0|TP=5|ip-26-0-174-100]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=54|PP=0|TP=6|ip-26-0-174-100]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=54|PP=0|TP=7|ip-26-0-174-100]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=54|PP=0|TP=3|ip-26-0-174-100]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=54|PP=0|TP=2|ip-26-0-174-100]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=14|PP=0|TP=4|ip-26-0-164-75]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=41|PP=0|TP=7|ip-26-0-171-230]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=17|PP=0|TP=5|ip-26-0-165-213]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=17|PP=0|TP=6|ip-26-0-165-213]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=17|PP=0|TP=3|ip-26-0-165-213]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=17|PP=0|TP=7|ip-26-0-165-213]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=22|PP=0|TP=4|ip-26-0-166-214]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=22|PP=0|TP=0|ip-26-0-166-214]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=5|PP=0|TP=0|ip-26-0-163-158]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=43|PP=0|TP=0|ip-26-0-171-56]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=43|PP=0|TP=2|ip-26-0-171-56]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=43|PP=0|TP=4|ip-26-0-171-56]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=43|PP=0|TP=5|ip-26-0-171-56]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=49|PP=0|TP=4|ip-26-0-172-57]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=49|PP=0|TP=2|ip-26-0-172-57]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=49|PP=0|TP=6|ip-26-0-172-57]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=29|PP=0|TP=5|ip-26-0-168-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=9|PP=0|TP=6|ip-26-0-164-0]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=2|PP=0|TP=6|ip-26-0-161-78]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=2|PP=0|TP=5|ip-26-0-161-78]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=2|PP=0|TP=0|ip-26-0-161-78]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=2|PP=0|TP=4|ip-26-0-161-78]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=26|PP=0|TP=5|ip-26-0-167-51]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=26|PP=0|TP=3|ip-26-0-167-51]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=26|PP=0|TP=1|ip-26-0-167-51]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=23|PP=0|TP=5|ip-26-0-166-244]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=23|PP=0|TP=4|ip-26-0-166-244]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=21|PP=0|TP=4|ip-26-0-166-15]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=21|PP=0|TP=3|ip-26-0-166-15]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=21|PP=0|TP=5|ip-26-0-166-15]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=21|PP=0|TP=0|ip-26-0-166-15]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=21|PP=0|TP=6|ip-26-0-166-15]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=59|PP=0|TP=6|ip-26-0-175-165]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=4|PP=0|TP=5|ip-26-0-162-46]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=63|PP=0|TP=4|ip-26-0-175-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=63|PP=0|TP=5|ip-26-0-175-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=63|PP=0|TP=2|ip-26-0-175-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=30|PP=0|TP=5|ip-26-0-168-95]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=11|PP=0|TP=2|ip-26-0-164-187]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=11|PP=0|TP=0|ip-26-0-164-187]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=38|PP=0|TP=5|ip-26-0-170-31]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=38|PP=0|TP=7|ip-26-0-170-31]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=53|PP=0|TP=1|ip-26-0-173-7]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=53|PP=0|TP=6|ip-26-0-173-7]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=53|PP=0|TP=5|ip-26-0-173-7]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=53|PP=0|TP=0|ip-26-0-173-7]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=53|PP=0|TP=3|ip-26-0-173-7]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=53|PP=0|TP=7|ip-26-0-173-7]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=33|PP=0|TP=1|ip-26-0-169-207]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=50|PP=0|TP=2|ip-26-0-172-73]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=50|PP=0|TP=0|ip-26-0-172-73]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=50|PP=0|TP=6|ip-26-0-172-73]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=50|PP=0|TP=5|ip-26-0-172-73]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=62|PP=0|TP=1|ip-26-0-175-241]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=62|PP=0|TP=2|ip-26-0-175-241]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=62|PP=0|TP=7|ip-26-0-175-241]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=62|PP=0|TP=4|ip-26-0-175-241]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=46|PP=0|TP=3|ip-26-0-172-116]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=46|PP=0|TP=0|ip-26-0-172-116]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=60|PP=0|TP=0|ip-26-0-175-170]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=60|PP=0|TP=2|ip-26-0-175-170]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=48|PP=0|TP=1|ip-26-0-172-147]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=37|PP=0|TP=0|ip-26-0-170-160]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=1|PP=0|TP=4|ip-26-0-161-123]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=1|PP=0|TP=6|ip-26-0-161-123]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=1|PP=0|TP=2|ip-26-0-161-123]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=1|PP=0|TP=7|ip-26-0-161-123]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=12|PP=0|TP=6|ip-26-0-164-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=45|PP=0|TP=2|ip-26-0-171-88]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=10|PP=0|TP=4|ip-26-0-164-18]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=10|PP=0|TP=1|ip-26-0-164-18]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=10|PP=0|TP=6|ip-26-0-164-18]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=28|PP=0|TP=3|ip-26-0-168-120]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=19|PP=0|TP=4|ip-26-0-165-38]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=19|PP=0|TP=3|ip-26-0-165-38]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=34|PP=0|TP=0|ip-26-0-169-247]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=34|PP=0|TP=2|ip-26-0-169-247]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=34|PP=0|TP=6|ip-26-0-169-247]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=34|PP=0|TP=3|ip-26-0-169-247]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=34|PP=0|TP=5|ip-26-0-169-247]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=27|PP=0|TP=2|ip-26-0-167-9]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=27|PP=0|TP=4|ip-26-0-167-9]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=27|PP=0|TP=5|ip-26-0-167-9]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=27|PP=0|TP=1|ip-26-0-167-9]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=35|PP=0|TP=1|ip-26-0-170-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=35|PP=0|TP=2|ip-26-0-170-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=35|PP=0|TP=3|ip-26-0-170-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=51|PP=0|TP=0|ip-26-0-173-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=51|PP=0|TP=5|ip-26-0-173-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=51|PP=0|TP=3|ip-26-0-173-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=51|PP=0|TP=4|ip-26-0-173-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=3|PP=0|TP=5|ip-26-0-162-233]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=3|PP=0|TP=6|ip-26-0-162-233]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=3|PP=0|TP=0|ip-26-0-162-233]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=3|PP=0|TP=7|ip-26-0-162-233]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=3|PP=0|TP=1|ip-26-0-162-233]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=55|PP=0|TP=0|ip-26-0-174-186]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=55|PP=0|TP=4|ip-26-0-174-186]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=36|PP=0|TP=2|ip-26-0-170-143]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=36|PP=0|TP=0|ip-26-0-170-143]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=40|PP=0|TP=0|ip-26-0-171-21]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=8|PP=0|TP=7|ip-26-0-163-43]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=8|PP=0|TP=5|ip-26-0-163-43]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=25|PP=0|TP=1|ip-26-0-167-177]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=42|PP=0|TP=6|ip-26-0-171-249]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=31|PP=0|TP=0|ip-26-0-169-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=31|PP=0|TP=3|ip-26-0-169-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=31|PP=0|TP=7|ip-26-0-169-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=13|PP=0|TP=7|ip-26-0-164-45]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=18|PP=0|TP=5|ip-26-0-165-24]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=18|PP=0|TP=6|ip-26-0-165-24]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=18|PP=0|TP=7|ip-26-0-165-24]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=52|PP=0|TP=2|ip-26-0-173-246]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=57|PP=0|TP=4|ip-26-0-174-240]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=57|PP=0|TP=6|ip-26-0-174-240]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=20|PP=0|TP=1|ip-26-0-165-59]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=39|PP=0|TP=2|ip-26-0-171-168]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=58|PP=0|TP=7|ip-26-0-175-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=24|PP=0|TP=1|ip-26-0-167-175]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=61|PP=0|TP=3|ip-26-0-175-19]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=61|PP=0|TP=5|ip-26-0-175-19]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=61|PP=0|TP=6|ip-26-0-175-19]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=61|PP=0|TP=0|ip-26-0-175-19]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=61|PP=0|TP=2|ip-26-0-175-19]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=15|PP=0|TP=4|ip-26-0-165-164]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=15|PP=0|TP=5|ip-26-0-165-164]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=7|PP=0|TP=7|ip-26-0-163-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=7|PP=0|TP=3|ip-26-0-163-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=7|PP=0|TP=2|ip-26-0-163-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=7|PP=0|TP=5|ip-26-0-163-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=54|PP=0|TP=4|ip-26-0-174-100]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=14|PP=0|TP=1|ip-26-0-164-75]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=14|PP=0|TP=0|ip-26-0-164-75]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=14|PP=0|TP=6|ip-26-0-164-75]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=14|PP=0|TP=5|ip-26-0-164-75]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=14|PP=0|TP=7|ip-26-0-164-75]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=41|PP=0|TP=3|ip-26-0-171-230]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=41|PP=0|TP=2|ip-26-0-171-230]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=41|PP=0|TP=5|ip-26-0-171-230]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=22|PP=0|TP=3|ip-26-0-166-214]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=22|PP=0|TP=7|ip-26-0-166-214]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=5|PP=0|TP=7|ip-26-0-163-158]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=5|PP=0|TP=3|ip-26-0-163-158]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=5|PP=0|TP=4|ip-26-0-163-158]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=5|PP=0|TP=5|ip-26-0-163-158]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=5|PP=0|TP=2|ip-26-0-163-158]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=43|PP=0|TP=3|ip-26-0-171-56]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=29|PP=0|TP=7|ip-26-0-168-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=29|PP=0|TP=3|ip-26-0-168-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=29|PP=0|TP=0|ip-26-0-168-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=29|PP=0|TP=6|ip-26-0-168-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=29|PP=0|TP=4|ip-26-0-168-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=9|PP=0|TP=0|ip-26-0-164-0]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=9|PP=0|TP=7|ip-26-0-164-0]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=9|PP=0|TP=1|ip-26-0-164-0]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=2|PP=0|TP=3|ip-26-0-161-78]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=2|PP=0|TP=1|ip-26-0-161-78]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=26|PP=0|TP=4|ip-26-0-167-51]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=26|PP=0|TP=7|ip-26-0-167-51]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=23|PP=0|TP=0|ip-26-0-166-244]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=23|PP=0|TP=1|ip-26-0-166-244]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=23|PP=0|TP=7|ip-26-0-166-244]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=23|PP=0|TP=2|ip-26-0-166-244]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=21|PP=0|TP=2|ip-26-0-166-15]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=59|PP=0|TP=3|ip-26-0-175-165]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=59|PP=0|TP=2|ip-26-0-175-165]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=59|PP=0|TP=1|ip-26-0-175-165]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=59|PP=0|TP=4|ip-26-0-175-165]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=63|PP=0|TP=6|ip-26-0-175-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=63|PP=0|TP=1|ip-26-0-175-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=30|PP=0|TP=3|ip-26-0-168-95]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=11|PP=0|TP=7|ip-26-0-164-187]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=38|PP=0|TP=0|ip-26-0-170-31]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=33|PP=0|TP=6|ip-26-0-169-207]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=50|PP=0|TP=7|ip-26-0-172-73]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=50|PP=0|TP=1|ip-26-0-172-73]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=62|PP=0|TP=6|ip-26-0-175-241]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=46|PP=0|TP=5|ip-26-0-172-116]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=46|PP=0|TP=6|ip-26-0-172-116]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=46|PP=0|TP=7|ip-26-0-172-116]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=46|PP=0|TP=2|ip-26-0-172-116]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=37|PP=0|TP=4|ip-26-0-170-160]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=1|PP=0|TP=0|ip-26-0-161-123]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=12|PP=0|TP=0|ip-26-0-164-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=12|PP=0|TP=1|ip-26-0-164-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=12|PP=0|TP=7|ip-26-0-164-236]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=45|PP=0|TP=4|ip-26-0-171-88]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=45|PP=0|TP=3|ip-26-0-171-88]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=45|PP=0|TP=5|ip-26-0-171-88]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=45|PP=0|TP=1|ip-26-0-171-88]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=10|PP=0|TP=2|ip-26-0-164-18]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=28|PP=0|TP=4|ip-26-0-168-120]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=19|PP=0|TP=6|ip-26-0-165-38]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=19|PP=0|TP=2|ip-26-0-165-38]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=19|PP=0|TP=7|ip-26-0-165-38]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=34|PP=0|TP=7|ip-26-0-169-247]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=27|PP=0|TP=7|ip-26-0-167-9]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=51|PP=0|TP=2|ip-26-0-173-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=51|PP=0|TP=7|ip-26-0-173-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=51|PP=0|TP=6|ip-26-0-173-202]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=3|PP=0|TP=2|ip-26-0-162-233]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=55|PP=0|TP=5|ip-26-0-174-186]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=55|PP=0|TP=1|ip-26-0-174-186]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=36|PP=0|TP=3|ip-26-0-170-143]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=25|PP=0|TP=2|ip-26-0-167-177]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=42|PP=0|TP=7|ip-26-0-171-249]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=42|PP=0|TP=2|ip-26-0-171-249]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=13|PP=0|TP=4|ip-26-0-164-45]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=18|PP=0|TP=3|ip-26-0-165-24]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=57|PP=0|TP=0|ip-26-0-174-240]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=39|PP=0|TP=7|ip-26-0-171-168]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=58|PP=0|TP=1|ip-26-0-175-132]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=24|PP=0|TP=0|ip-26-0-167-175]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=61|PP=0|TP=4|ip-26-0-175-19]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=15|PP=0|TP=7|ip-26-0-165-164]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=14|PP=0|TP=3|ip-26-0-164-75]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=41|PP=0|TP=0|ip-26-0-171-230]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=22|PP=0|TP=2|ip-26-0-166-214]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=5|PP=0|TP=1|ip-26-0-163-158]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=29|PP=0|TP=1|ip-26-0-168-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=9|PP=0|TP=4|ip-26-0-164-0]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=2|PP=0|TP=2|ip-26-0-161-78]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=59|PP=0|TP=0|ip-26-0-175-165]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=63|PP=0|TP=0|ip-26-0-175-34]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=30|PP=0|TP=0|ip-26-0-168-95]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=38|PP=0|TP=3|ip-26-0-170-31]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=62|PP=0|TP=5|ip-26-0-175-241]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=62|PP=0|TP=0|ip-26-0-175-241]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=37|PP=0|TP=6|ip-26-0-170-160]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=1|PP=0|TP=1|ip-26-0-161-123]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=45|PP=0|TP=0|ip-26-0-171-88]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=10|PP=0|TP=5|ip-26-0-164-18]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=28|PP=0|TP=7|ip-26-0-168-120]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=19|PP=0|TP=0|ip-26-0-165-38]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=19|PP=0|TP=1|ip-26-0-165-38]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=34|PP=0|TP=1|ip-26-0-169-247]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=27|PP=0|TP=0|ip-26-0-167-9]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=3|PP=0|TP=3|ip-26-0-162-233]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=55|PP=0|TP=2|ip-26-0-174-186]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=13|PP=0|TP=6|ip-26-0-164-45]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=18|PP=0|TP=2|ip-26-0-165-24]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=57|PP=0|TP=1|ip-26-0-174-240]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=39|PP=0|TP=0|ip-26-0-171-168]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=41|PP=0|TP=4|ip-26-0-171-230]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=41|PP=0|TP=6|ip-26-0-171-230]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=9|PP=0|TP=5|ip-26-0-164-0]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=30|PP=0|TP=1|ip-26-0-168-95]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=38|PP=0|TP=1|ip-26-0-170-31]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=37|PP=0|TP=1|ip-26-0-170-160]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=10|PP=0|TP=7|ip-26-0-164-18]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=55|PP=0|TP=7|ip-26-0-174-186]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=13|PP=0|TP=0|ip-26-0-164-45]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=1|PP=0|TP=3|ip-26-0-161-123]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=10|PP=0|TP=0|ip-26-0-164-18]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=37|PP=0|TP=2|ip-26-0-170-160]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=44|PP=0|TP=7|ip-26-0-171-62]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=44|PP=0|TP=0|ip-26-0-171-62]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=44|PP=0|TP=4|ip-26-0-171-62]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=44|PP=0|TP=5|ip-26-0-171-62]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=44|PP=0|TP=6|ip-26-0-171-62]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=44|PP=0|TP=3|ip-26-0-171-62]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=44|PP=0|TP=2|ip-26-0-171-62]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=6|PP=0|TP=1|ip-26-0-163-220]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=6|PP=0|TP=2|ip-26-0-163-220]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=6|PP=0|TP=3|ip-26-0-163-220]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=6|PP=0|TP=6|ip-26-0-163-220]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=6|PP=0|TP=7|ip-26-0-163-220]: Throughput logging complete +11/27/2024 14:26:05 [INFO|DP=6|PP=0|TP=0|ip-26-0-163-220]: Throughput logging complete +[2024-11-27 14:26:05,131] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,131] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23931 closing signal SIGTERM +[2024-11-27 14:26:05,132] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23932 closing signal SIGTERM +[2024-11-27 14:26:05,132] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,133] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,132] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23933 closing signal SIGTERM +[2024-11-27 14:26:05,132] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23934 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1166685 closing signal SIGTERM +[2024-11-27 14:26:05,132] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2808081 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2808082 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1166686 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2808083 closing signal SIGTERM +[2024-11-27 14:26:05,132] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1166687 closing signal SIGTERM +[2024-11-27 14:26:05,131] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2808084 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1310266 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1166688 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1310267 closing signal SIGTERM +[2024-11-27 14:26:05,131] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 448519 closing signal SIGTERM +[2024-11-27 14:26:05,131] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 448520 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1310268 closing signal SIGTERM +[2024-11-27 14:26:05,132] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,131] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 448521 closing signal SIGTERM +[2024-11-27 14:26:05,131] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 448522 closing signal SIGTERM +[2024-11-27 14:26:05,132] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 657199 closing signal SIGTERM +[2024-11-27 14:26:05,132] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 657200 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,132] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 657201 closing signal SIGTERM +[2024-11-27 14:26:05,132] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 657202 closing signal SIGTERM +[2024-11-27 14:26:05,132] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23935 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3341868 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3341869 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3341870 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83955 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83956 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83957 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1310269 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2808085 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 814830 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2808086 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2808087 closing signal SIGTERM +slurmstepd: error: *** JOB 11600246 ON ip-26-0-160-103 CANCELLED AT 2024-11-27T14:26:05 *** +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1166689 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 814831 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 814832 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 814833 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2828524 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2828525 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2828526 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 944721 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +slurmstepd: error: *** STEP 11600246.0 ON ip-26-0-160-103 CANCELLED AT 2024-11-27T14:26:05 *** +[2024-11-27 14:26:05,134] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2121107 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3023826 closing signal SIGTERM +[2024-11-27 14:26:05,132] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 448523 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 944722 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2121108 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3698367 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3023827 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1310270 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 944723 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2121109 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1107984 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3698368 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3023828 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2288489 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1107985 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1107986 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3698369 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3023829 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2288490 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3698370 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2288491 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1107987 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3046023 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83958 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3046024 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3023830 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 657203 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83959 closing signal SIGTERM +[2024-11-27 14:26:05,133] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 448524 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2828527 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2121110 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,134] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1151909 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 886345 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1151910 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4194122 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 886346 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4194123 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1151911 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1151912 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 23936 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4194124 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1310271 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 91587 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 4194125 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1310272 closing signal SIGTERM +[2024-11-27 14:26:05,134] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 886347 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 886348 closing signal SIGTERM +[2024-11-27 14:26:05,136] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 944724 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 83960 closing signal SIGTERM +[2024-11-27 14:26:05,136] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1166690 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 2121111 closing signal SIGTERM +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 3023831 closing signal SIGTERM +[2024-11-27 14:26:05,136] torch.distributed.elastic.agent.server.api: [WARNING] Received Signals.SIGTERM death signal, shutting down workers +[2024-11-27 14:26:05,135] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1107988 closing signal SIGTERM +[2024-11-27 14:26:05,136] torch.distributed.elastic.multiprocessing.api: [WARNING] Sending process 1107989 closing signal SIGTERM